Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Underwriting Sensitivity to Inflation and Claims Cost Shifts

Abstract. Underwriting sensitivity to inflation is the responsiveness of prospective insurance cost, contract performance, portfolio composition, and capital requirements to changes in the prices and quantities of resources consumed by insured events. A common price index provides insufficient information for determining the sensitivity of an individual contract because claim settlement is governed by heterogeneous cost components, payment delays, nonproportional coverage, and changing selection into the portfolio. This monograph develops an international comparative synthesis of actuarial trending, claims development, sectoral economic measurement, contract mechanics, renewal decisions, and prudential governance. It examines repair materials and labor, medical expenditure, legal and settlement costs, supply-chain disruption, and exposure valuation across property, motor, casualty, and related non-life insurance. The analysis combines published institutional and scholarly evidence with explicitly hypothetical, reproducible sensitivity calculations. It demonstrates how a fixed deductible can magnify percentage changes in covered severity, how a payment limit can attenuate them, how long settlement duration compounds inflation uncertainty, and how apparently adequate rate changes can lose effectiveness through earning lags, exposure changes, and deteriorating business mix. A common inflation scenario also creates dependence across otherwise diversified exposures. The central conclusion is that underwriting adequacy requires prospective, coverage-specific, time-consistent scenarios that coordinate risk selection, contractual terms, technical pricing, reinsurance, and financial planning. The analysis makes no claim to a new theoretical system or to estimation from confidential insurer records.

Keywords: Insurance underwriting; claims inflation; loss severity; exposure valuation; actuarial trending; social inflation; deductibles; policy limits; premium adequacy; reinsurance; renewal underwriting; insurance capital; portfolio sensitivity.

Introduction and disciplinary foundation

The object and significance of underwriting sensitivity

Underwriting is a prospective decision about the acceptance of a risk and the financial and contractual conditions under which that risk can enter an insurance portfolio. Inflation changes the relationship between those conditions and the future cost of honoring the resulting promises. The object of inquiry is therefore the economic sensitivity of an underwriting decision over the interval separating the evaluation of an exposure from the ultimate fulfillment of the associated obligations.

An underwriter ordinarily observes the characteristics of an exposure before its losses have occurred. A claims handler observes a realized event and determines the compensation due under a legally operative contract. The actuary estimates distributions and develops assumptions that connect evidence from past events to future obligations. All three functions encounter price changes, but their decisions differ. A claims reserve concerns liabilities generated by events or coverage already recognized under the applicable accounting or prudential framework; prospective technical cost concerns the obligations expected from future coverage. Underwriting coordinates that expectation with terms, eligibility, and capacity. Keeping these objects distinct prevents a mechanical transfer of reserving assumptions into a renewal decision.

Inflation belongs inside the expected-loss distribution. The cost of restoring a damaged structure reflects construction labor, replacement materials, contractor capacity, code requirements, and the time necessary to complete restoration. The cost of a bodily injury includes medical services, future care, lost earnings where recoverable, and legal expenses. A liability settlement may reflect court practices that evolve differently from consumer prices. An underwriting system that uses one scalar inflation assumption for all these costs risks serious aggregation error even when that assumption is a plausible description of the general economy.

The experience documented in EIOPA's 2023 report on inflation illustrates the practical significance. Its aggregate EEA non-life comparison from the fourth quarter of 2021 to the fourth quarter of 2022 reported incurred claims rising 11.3%, expenses 4.3%, and earned premiums 6.4%, with the combined ratio rising from 97.5% to 99.9%. These are aggregate period comparisons within the report's specified reporting population. Changes in coverage volume, portfolio mix, reserving and exchange rates can affect them; the 11.3% observation cannot be interpreted as a universal per-claim inflation rate. The comparison nonetheless establishes why differences in the timing of premium growth and claims-cost growth matter economically.

The underwriting question is not simply whether prices are increasing. It is which future cash flows are sensitive to which prices, over what horizons, with what contractual response and under what information and institutional constraints. An insurer can face a material deterioration in underwriting economics while headline inflation falls, because earlier increases remain embedded in procurement costs, wage agreements, and outstanding repairs. Conversely, an increase in a broad index need not produce proportional increases in a particular insurer's expected payout where limits, substitutions, or shifts in frequency act in the opposite direction.

Research questions and methodological design

Three questions organize this investigation. First, how do inflation pass-through and changes in claim composition alter prospective loss distributions? Second, how should cost indices be selected for different insurance classes and durations? Third, when are adjustments in limits, deductibles, renewal terms and risk appetite economically justified and institutionally feasible?

The analysis is a documentary and mathematical synthesis. It uses international supervisory principles, professional standards, official statistical methodologies, dated supervisory reports, comparative market observations, and selected scholarly research. Contemporary evidence is deliberately attributed to its jurisdiction and period. Published research summaries and abstracts support only the specific findings visible in those records. The research cutoff is November 19, 2026. The numerical illustrations are deterministic scenarios with specified distributions and arithmetic; they are not predictions or fitted empirical models. No confidential insurance data, actual contract file or policyholder record has been used.

The core scientific domain is property/casualty underwriting, with medical and labor-cost evidence incorporated wherever it affects covered non-life losses. Long-duration life and health products have distinct benefit structures, morbidity, mortality, guarantees and regulatory arrangements; the general cash-flow reasoning can inform them, but the numerical examples and institutional conclusions here do not establish product-specific pricing or legal rules for those markets.

Time, measurement and the unit of analysis

A defensible analysis must specify the unit whose sensitivity is measured. An insured location, vehicle-year, liability contract, claim, paid transaction and annual portfolio dollar are different denominators. Claims cost per vehicle-year can rise because the number of accidents per vehicle-year increases, average covered payments increase, or exposure composition changes. Claims cost per reported claim can rise when minor claims are no longer reported. An increase in total paid claims can follow portfolio growth without any deterioration in cost per unit of risk.

Time likewise has multiple meanings. Policy inception dates locate the contractual promise; accident dates locate the event; reporting dates locate the information arrival; settlement dates locate the discharge; payment dates locate cash consumption. An estimate that combines accident-year frequency with calendar-year paid severity must account for these different populations. The economic question underlying each decision determines the relevant observation period.

The appropriate comparison begins with a consistent exposure unit and coverage specification, separates observed from forecast periods, distinguishes nominal from real movements, and identifies which obligations respond to costs as they occur rather than costs when a claim was reported. These distinctions are central to the interpretation of all examples that follow.

Claims cost inflation mechanisms

Frequency, severity and cost per exposure

For a homogeneous exposure unit, let N denote the number of claims, X an individual claim payment, and S aggregate claims cost. Under the conventional conditional-independent collective-risk illustration, expected aggregate cost is E[N] times E[X]. More generally, the conditional distribution of claim amounts can vary with claim count and scenario, and the expectation must integrate those dependencies.

When frequency and severity are defined on the same population, expected cost per exposure is the product of claims per exposure and mean payout per claim. If frequency rises by 2% and mean severity rises by 8%, the multiplicative increase is 1.02 × 1.08 − 1 = 10.16%. An additive approximation of 10% is close but omits the interaction. The economic interpretation changes if frequency is measured after a deductible or only after claims exceed a reporting threshold. A rising nominal cost level can push previously uncovered events above a fixed threshold and thereby change observed claim frequency even if physical event frequency is stable.

Claims-cost inflation may arise from input prices, changes in quantities used per claim, treatment or repair techniques, damages valuation, settlement behavior, and the distribution of event types. A labor rate increase differs from an increase in labor hours required to repair a particular vehicle. An expensive electronic component replacing a simple mechanical component changes the bill of materials. Supply constraints can increase both the price of a part and the time for which a replacement vehicle is required. These mechanisms produce different risk-management responses.

The ASB's trending procedures standard (2009) distinguishes the projection of costs from one period to another from the development of the estimate for an existing exposure period. This distinction supplies a methodological foundation. A historical incurred-to-ultimate factor may incorporate past inflation, claim closure practices and case-reserving behavior. Applying a second undifferentiated inflation factor to the same historical development creates a double-counting risk. Conversely, the use of historical development factors in a prospective projection without consideration of a changed future inflation regime can embed obsolete assumptions.

Calendar effects and identification

Claims triangles summarize data by origin period and development age. Calendar-period influences occur diagonally. Taylor's (1977) separation of inflation effects addresses the historical actuarial problem of distinguishing inflation from other influences on claims development. A sustained change in settlement cost affects multiple open accident cohorts, but the pattern also reflects the settlement process.

An illustrative log-linear representation for positive expected payments is a_i + b_j + c_t, where a_i is an origin effect, b_j a development effect, c_t a calendar effect, and t = i + j under consistent integer indexing. Adding k i to the origin effect, k j to the development effect and subtracting k t from the calendar effect leaves the linear predictor unchanged. This identity establishes an identification problem: the three unconstrained linear trends cannot be recovered uniquely from the same triangle. Normalizing intercepts alone does not eliminate the linear-drift ambiguity.

The analyst can constrain the model with relevant external price indexes, independently observed claim counts, documented settlement-process changes, or explicit assumptions about trend shape. Each such choice affects inference. The external price series is evidence about an economic driver, not proof that every diagonal movement arises from that price series. Case reserve adequacy, claims staffing, settlement incentives and reporting delays can produce competing explanations.

The synthetic examples in Creedon et al.'s (2025) claims inflation estimation guide illustrate the sensitivity of historical inference to method, thresholds and frequency. A robust underwriting process uses such diagnostic results to investigate whether apparently rising severity persists when coverage, claims thresholds, development age and risk composition are held comparable. It then carries identification uncertainty into forward scenarios instead of replacing it with a falsely precise point estimate.

Economic inflation, excess claims inflation and social inflation

Economic inflation describes changes in the price level of goods or services in a defined economic population. Claims inflation describes the change in an insurance loss measure after explicit choices about exposure, claim population, coverage and time. Excess claims inflation can be defined for operational analysis as the residual after an appropriate economic cost adjustment and specified composition controls. The residual is dependent on the chosen benchmark.

Social inflation is a term used in insurance research to describe non-economic or above-economic increases in liability claims costs associated with legal, institutional and social changes. Swiss Re Institute's 2024 social inflation study relates this issue to litigation costs, particularly in US liability insurance. Its findings are industry estimates with the definitions and market scope of that report. Lynch and Moore's (2023) loss development update reports an estimated impact exceeding $30 billion on US commercial auto liability claims over 2012–2021. This is modeled attribution to an identified market and period, not an experimentally established global causal effect.

The distinction between directly observed claim payments and inferred explanatory factors is consequential. Higher awards, changing venue composition, case selection, policy limits, claim size, and judicial delay can interact. A residual above consumer inflation also includes omitted economic variables, changing case mix, model error and possible measurement changes. Thus, a residual series is useful as a diagnostic and stress input, but a claim about a specific legal mechanism requires separate evidence. Estimating a portable social inflation coefficient from a single market summary is scientifically unwarranted.

Underwriters should distinguish attritional legal-cost escalation from the probability of exceptionally large awards and from the allocation of defense expenses under contract wording. A mean-severity adjustment may understate the economic effect when the probability of large losses is increasing. Long-tail excess coverage can be especially sensitive to an upward shift in the upper tail even before the mean of paid claims identifies the change.

Supply disruption, catastrophe demand surge and persistent change

An insured loss creates demand for repair goods or services. When many losses occur simultaneously, the relevant local market can experience labor and equipment shortages. The resulting cost increase is conditional on the event and location. Treating catastrophe-related demand surge as independent general inflation may distort both expected losses and portfolio tail risk.

Supply disruption also changes claim duration, availability of substitutes, and the composition of repairs. Some costs are temporary, others persist because nominal wages or service agreements adjust with a lag. A prospective scenario should distinguish a price-level step from an ongoing annual growth rate. A one-time 15% increase in parts prices that subsequently remain stable requires a level adjustment; compounding 15% every year extrapolates a shock into a new trend without supporting evidence.

The risk of inflation persistence is related to the contractual and procurement environment. Materials may be purchased at a fixed price shortly after an event; labor may be paid over several months; future medical services may extend over years. An appropriate cost scenario specifies when prices enter actual settlement amounts. It also captures the possibility that changes in procurement practice, early settlement, prevention or repair substitution alter the pass-through.

Sector-specific indices

Index selection as a measurement problem

A price index measures the movement of a defined basket according to a specified methodology. The correct choice depends on the actual costs underlying the insured obligations. Consumer price indexes describe household consumption under defined statistical conventions. Producer and input price measures capture different transaction stages. Wage measures capture labor compensation. Medical expenditure measures may combine prices, utilization and changes in services.

A weighted cost basket can be formed from estimated baseline shares of repair materials, labor, professional fees, medical treatment and other costs. For fixed quantities, component price relatives can be combined with their baseline shares to obtain a scenario price relative. However, the exposure's basket can itself change over time. An estimate based on fixed weights isolates a price component only to the extent that the quantities and quality definitions remain comparable.

For example, a hypothetical repair consists of 55% materials, 35% labor and 10% overhead. If the three cost categories rise 12%, 6% and 4%, respectively, the fixed-basket increase is 0.55 × 0.12 + 0.35 × 0.06 + 0.10 × 0.04 = 9.1%. This is a gross repair-cost increase before coverage terms, claim mix and changing quantities. A corresponding headline inflation rate is irrelevant unless it has been established as a reliable approximation for these specific inputs.

The BLS (2022) industry input price indexes distinguish selected purchased inputs from labor and capital costs. For US labor, the BLS (2025) Employment Cost Index measures changes in compensation using a fixed composition of occupations and industries. Its wage and benefit coverage is useful when assessing compensation pressure, but actual contractor hourly charges can also include overhead, productivity, profit margins and local shortages.

Property and construction exposures

For insured buildings, cost replacement is often embedded in the insured value and the projected claim severity. Those quantities are related but serve different purposes. An updated building valuation changes the size of the exposure or adequacy of the limit; it does not by itself specify the cost distribution of the events that generate claims.

Useful index candidates include construction material prices, local labor costs, equipment rental, and sector-specific rebuilding estimates. The relevant weights depend on occupancy, quality, location, size and building systems. A commercial warehouse and a historic masonry building can face different cost pressures even in the same city. Reconstruction after a covered event may also entail demolition, waste disposal, code compliance, engineering services, and extraordinary demand on contractors.

Index-based valuation adjustments can improve the alignment of sums insured with replacement values across a portfolio. They remain sensitive to baseline valuation accuracy and geographic granularity. Properties with substantial subsequent renovations, changes in occupancy or nonstandard construction require individualized assessment. An index should update a valid valuation, not serve as a substitute for establishing its initial level.

Business interruption creates another mechanism. The duration of inability to operate may rise because replacement parts or specialist contractors are scarce. The insured financial loss then responds to both the economics of the interrupted business and the duration of disruption. An index of construction prices alone misses this time component. Underwriting should evaluate the contract's indemnity period, waiting period, declared values, dependencies and realistic restoration plans.

Motor and repair-sensitive insurance

Motor claims involve physical repair, parts, labor, replacement vehicles, towing, storage, total loss settlements and bodily injury where covered. Vehicle technology can alter the resources required to repair otherwise comparable damage. A shift from repairable to total-loss outcomes may have nonlinear effects because the replacement value and salvage recovery jointly determine loss.

An analyst should separate repair cost per repairable claim from total-loss value, claim duration, and the changing proportion of total losses. Price movements in used vehicles can change the relative economic attractiveness of repair versus replacement. A unified motor severity series may obscure this tradeoff, especially after a shift in policy limits or deductibles.

Bohnert, Gatzert and Kolb's (2016) non-life inflation risk study investigates class-dependent inflation drivers. The authors' 2015 prepublication research manuscript describes the setting of a large German insurer and distinguishes motor liability, motor comprehensive and third-party liability. The study supports the relevance of line-specific empirical investigation. Its particular market and insurer setting limits generalization; the research record and prepublication version do not justify treating reported coefficients as directly transferable to another portfolio.

Medical, wage and bodily injury components

Medical costs must be decomposed into unit prices, utilization, service composition, duration, and the proportion payable under the coverage. Medical inflation may arise in physician services, inpatient care, rehabilitation, prescription medicines and long-term assistance. Actual liability payouts can reflect jurisdiction-specific damages rules and evidentiary requirements.

The BLS medical CPI methodology clarifies consumer-expenditure scope and treatment of medical care components. Its measure is not identical to the negotiated allowed amounts of a particular insurer, or to the future care costs assessed in an injury case. CMS (2026) national health expenditure accounts report total US health spending of $5.3 trillion in 2024, an increase of 7.2%. The spending measure combines price, use and composition changes across payers and is unsuitable as a direct estimate of a particular insurance claim's price trend.

For recoverable income loss, compensation changes may matter separately from healthcare cost trends. For future assistance, the wage costs of care workers, availability of services and duration of the award may matter more than consumer medical prices. The best index for each part of a claim is the index whose transaction population and price concept align with that part.

An underwriting forecast should distinguish benefit schedules indexed by statute or contract from indemnities determined by actual economic loss. A legally fixed benefit changes only through the mechanism applicable to it, while a cost-based reimbursement can respond more continuously to changes in service prices. The same healthcare shock can therefore have radically different effects across products.

Legal expenses, liability and supply-chain indicators

Litigation-related costs require analysis of external legal fees, expert testimony, defense expenses, settlement amounts and damages. Law-firm fee schedules can provide an input cost indicator, while changes in litigation probability or award composition require claims-process evidence. Case data should be analyzed with awareness of selection: settled and litigated cases are different populations.

For supply chains, Benigno et al.'s (2022) global supply chain pressure index offers a macroeconomic indicator of disruption that can inform directional scenarios. Such an index is a predictor candidate, not a direct price index for covered losses. A model using it needs a defensible lag structure, segment-specific validation and evidence that the relevant insurer's contracts are sensitive to those disruptions.

Exchange rates can also alter the domestic currency cost of imported components. A contract denominated in one currency may cover replacement of an item purchased internationally in another. Local wage and material indexes cannot capture all currency effects. The forecast should specify the currency in which the loss is payable, the currency of underlying procurement, and the degree of contractual or operational hedging.

From index to insured cost distribution

A price series should be matched to the appropriate segment, cost component, geography, currency, release frequency and lag. Forecast scenarios can combine prospective paths for component indexes with changes in the quantities used. The resulting ground-up loss distribution is then transformed by contract terms. This sequence protects against applying a price increase directly to insurer payouts when those payouts are affected by limits or deductibles.

A data register should retain the precise index name, statistical agency, series identification where used, vintage, reference period, seasonal adjustment, weights, currency and revision policy. Backtesting should use the data that would have been available when the original underwriting decision was made. A revised historical series can be useful for retrospective diagnosis but may introduce look-ahead bias when used to simulate decision quality.

The selection of an index is itself a source of model risk. Multiple correlated series may predict similar outcomes in stable periods and diverge under shocks. A plausible forecast should be accompanied by competing scenarios and reasons for accepting or rejecting their underlying mechanisms.

Exposure and limit adjustments

Ground-up loss and contractual transformation

Insurance indemnity is a function of loss and applicable coverage. Consider an illustrative occurrence contract with ground-up loss X, monetary deductible d, and maximum insurer payment l. Ignoring other clauses, its insurer payment is Y = min(max(X − d, 0), l). This convention defines l as the maximum payment after the deductible; alternative policy wordings can place the limit on ground-up loss, apply coinsurance or use other orderings.

Let the loss-price scale be m. Then expected payout is G(m) = E[min(max(mX − d, 0), l)]. For a continuous distribution with finite mean and no mass at moving thresholds, its derivative is E[X 1{d < mX < d + l}]. The derivative expresses the local exposure of expected payout to a proportional ground-up loss shift. Lower losses remain below the deductible, middle losses respond to the scale, and losses already exhausting the maximum payment are capped. For distributions with atoms at a boundary, directional changes require direct evaluation.

The function is nonlinear. Multiplying the expected payout by the headline inflation factor is exact only in special cases, such as proportional scaling of all relevant monetary thresholds alongside losses under the simplified contract. A portfolio with multiple deductibles and limits requires application at the appropriate contract or risk level before aggregation.

A reproducible deductible and limit experiment

Consider an invented three-point loss distribution. Ground-up loss is $500 with probability 0.50, $2,000 with probability 0.40, and $10,000 with probability 0.10. Assume one event per observation; zero claims are outside the illustration. The fixed deductible is $1,000 and the maximum insurer payment is $5,000. Increase every ground-up loss by 20% while holding the probabilities fixed.

Configuration Baseline expected payout Shock expected payout Change
Ground-up loss $2,050 $2,460 20.00%
Fixed deductible and payment cap $900 $1,060 17.78%
Fixed deductible, no cap $1,300 $1,660 27.69%
Both monetary thresholds indexed by 20% $900 $1,080 20.00%

The capped baseline payout is 0.40 × $1,000 + 0.10 × $5,000 = $900. With the shock, the middle loss pays $1,400, while the large loss remains capped. The resulting expected payout is $1,060. Removing the cap exposes more of the underlying increase; indexing both thresholds preserves proportionality in this particular example.

The deductible can magnify percentage changes in covered severity because the deductible subtracts a fixed amount from a growing loss. The cap counteracts this effect by excluding additional loss above the payment maximum. Their joint effect depends on the distribution. A $900 loss rising to $1,080 would cross the fixed $1,000 deductible and create an $80 payment; thus even observed positive-payment frequency can respond to nominal price changes. This is a contract threshold effect, not evidence of changing physical hazard.

The consumer's retained burden is also sensitive. A fixed limit can produce increasing uninsured loss as reconstruction values rise. An increased deductible shifts the initial layer of loss to the policyholder. The evaluation of a renewal offer must consider financial adequacy to the insurer and meaningful protection for the policyholder, along with applicable legal requirements.

Exposure denominators, valuation and adequacy

Exposure measurement determines how risk costs are expressed. A property rate per thousand dollars of insured value can remain unchanged while the total charged premium rises as the insured value increases. A rate increase layered on top of valuation growth can produce larger total premium growth than either adjustment considered alone. Conversely, a stale declared value may create both underinsurance and distortions in technical rate adequacy.

Consider an insured value rising by 12% and a rate per dollar of value rising by 5%. Holding all other factors fixed, premium rises by 1.12 × 1.05 − 1 = 17.6%. Treating the 17.6% premium change as a pure rate increase would overstate the pricing action. Treating the 12% valuation change as a complete measure of loss cost adjustment would ignore differences between the exposure denominator and the relevant claim cost dynamics.

A coherent renewal comparison reports both changes separately and specifies the expected claim cost per unit of exposure. It also distinguishes the policy limit from the value of the exposed property: a limit lower than replacement value can produce a capped insurer obligation, but the uncovered portion may make protection less effective. Contractual coinsurance penalties and valuation rules must be evaluated according to the actual wording and governing law.

For liability, exposure bases can include revenue, payroll, units produced or other activity indicators. Nominal payroll can rise through wage inflation even with constant hours worked. An insurance premium based on that payroll may increase automatically, while the underlying exposure to accidents and the severity of bodily injury follow different mechanisms. Normalizing loss costs using an inflation-sensitive denominator can conceal growing absolute losses unless both numerator and denominator are reconciled.

Limits, attachment points and excess insurance

Excess layers respond to shifts in the upper part of a loss distribution. An increase in all losses can move additional observations across attachment points and increase exhaustion of layers. This effect is distinct from a proportional change in a primary insurance loss ratio.

For illustration, suppose three separate covered events have losses of $0.7 million, $1.5 million and $3.5 million. An occurrence reinsurance layer pays up to $2 million above $1 million. Under simplified terms, the three ceded losses total $2.5 million, and the retained losses total $3.2 million. With a 20% ground-up shock, gross losses become $0.84 million, $1.8 million and $4.2 million. Ceded losses total $2.8 million and retained losses $4.04 million. Gross loss increases 20%, while retained loss increases 26.25%.

The result depends on the assumed discrete events and caps; alternative distributions and treaty structures can increase ceded sensitivity instead. The example excludes premium changes, reinstatement costs, aggregate limits, exclusions, timing and counterparty effects. It demonstrates why ground-up inflation must be translated through the actual risk transfer structure before measuring net portfolio sensitivity.

A higher attachment can reduce cessions and retain more frequent or larger losses. A higher layer limit can increase transfer of the tail but may be unavailable or priced at a materially higher rate. Underwriting authorities need to examine contract changes in conjunction with the reinsurance program, rather than assume that a higher primary limit has no consequence for net accumulation.

Contract renewal dynamics

Timing and the earning mechanism

Underwriting action operates under a calendar of quotation, binding, inception, renewal and premium earning. A cost shock may affect settlement before it can be reflected in contracts eligible for renewal. Even after a rate change is implemented, its effect on earned premium emerges gradually.

Suppose an annual book has uniform inception dates, stable exposure, unchanged terms, no churn, and a 10% increase in rates applied to all new and renewing policies from the beginning of a year. If premium is earned uniformly over each policy term, the fraction of the earning stream at the new rate grows linearly from zero to one across that year. The average first-year fraction is one-half. The approximate first-year earned premium increase attributable to rate is thus 5%, while policies written during that year carry the full new rate. The calculation is a simplified continuous-time earning illustration, not a forecast of reported premium.

Seasonal exposure, cancellation, midterm adjustments, multi-year contracts, nonuniform earning and changes in composition all alter the comparison. It remains essential to separate written premium, earned premium and the rate charged for a standardized exposure. An underwriting report that celebrates a large written premium increase before examining the earned effect and current loss horizon can overstate protection against current cost pressure.

The Lloyd's underwriting profitability principle places technical pricing and portfolio management within managing agents' oversight framework. Its institutional significance is the integration of the plan and realized underwriting performance. It is a market-specific principle rather than a universal law, but the relationship between prospective assumptions and portfolio execution applies across insurance systems.

Selection, retention and changing composition

Price changes can alter the mix of business that renews. A portfolio average can deteriorate even when every fixed risk class experiences only modest inflation. Suppose a portfolio consists of ordinary and higher-severity claims. At baseline, the probabilities of the groups among claims are 75% and 25%; their mean payouts are $1,000 and $4,000. The mean is $1,750. If both severity means increase 5% and the claim mix remains fixed, mean payout becomes $1,837.50.

If the higher-severity share subsequently increases to 35%, the mean becomes 0.65 × $1,050 + 0.35 × $4,200 = $2,152.50, a 23% increase over baseline. The fixed-mix increase was 5%; the additional $315 is the composition effect evaluated at the new group means. This ordering defines one transparent decomposition; another ordering allocates interactions differently. No causal connection between rate changes and the hypothetical mix shift is assumed.

Real renewal data should compare eligible, quoted, offered, accepted, declined and nonrenewed exposures separately. Retention can be affected by affordability, competitor prices, distribution arrangements, credit constraints and risk appetite. Lower-risk customers may find better alternatives; alternatively, higher-risk customers may be more price sensitive. The direction of selection is an empirical question.

An insurer must distinguish the incidence of risk from the incidence of becoming a customer. A before-and-after severity comparison in the bound portfolio contains information about the composition of accepted business. It cannot identify the counterfactual experience of policies that were not bound. Stable segmentation and documentation of coverage changes support interpretation, but they do not recreate unobserved outcomes.

Rate actions, coverage changes and legal constraints

Underwriting can respond to forecast cost changes through technical rate, deductibles, limits, sublimits, endorsements, exposure declarations, risk-improvement measures, and capacity allocation. Each changes a different part of the economic relationship. A higher deductible transfers frequent retained losses; a lower limit transfers tail responsibility; a higher premium changes the payment for coverage; a prevention requirement may change the hazard itself. The actions are not economically interchangeable.

The permissible response is jurisdiction- and contract-specific. In the UK, FCA (2022) renewal pricing rules in ICOBS 6B.2 address covered home and motor insurance renewal prices relative to the equivalent new-business price. This rule limits a particular pricing relationship rather than freezing rates generally. Its scope and definitions matter; it cannot be extended automatically to all lines or jurisdictions.

In the United States, the NAIC's model law catalog illustrates that rate regulation and filing frameworks differ. A model text is a proposed standard for adoption, not itself binding state law. An insurer must determine the law and filing obligations applicable to the product and state, including when revised terms or rates may take effect. This monograph establishes no jurisdiction-specific filing entitlement.

Contract changes ordinarily require consideration of policy terms, endorsements, disclosure, renewal notices, and applicable policyholder protections. The scientific implication is that a forecasted premium adequacy gap and the legal ability to close that gap are separate variables. Where immediate adjustment is constrained, the institution can assess lawful alternatives, timing, reinsurance and changes in prospective appetite. The economic desirability of an action does not establish permission to implement it.

Governance of the renewal cycle

An effective feedback process distinguishes indicators that precede loss deterioration from outcomes that confirm it. Supplier quotations, wage agreements and claim-duration changes may provide early signals; paid severity, case reserve revisions and fully developed loss experience arrive at different lags. Each source can be informative, provided the timing and population are correctly specified.

Assessment can be organized by underwriting cohorts sharing policy effective periods, hazard classes, and comparable terms. Each cohort has a technical cost forecast, a set of accepted risks and a subsequent outcome. Forecast errors can be decomposed provisionally into exposure change, frequency, ground-up severity, contract transformation, composition and development. The decomposition is an analytical aid; the attribution of interacting effects depends on the calculation order and evidence.

Monitoring should include retention, exposure-adjusted rate change, total premium change, coverage adequacy, accepted business mix, and changes in tail exposure. The use of thresholds for review should reflect model uncertainty, speed of deterioration and decision costs. A stable monthly estimate can still support timely escalation where a new supply or legal event materially affects a significant segment.

Economic and prudential sensitivity

Prospective premium adequacy

The prospective technical cost of insurance comprises more than expected indemnity. The ASB (2017) prospective cost estimation standard identifies categories including loss and loss adjustment expenses, operational expenses, reinsurance and capital costs as relevant to the estimation of future property/casualty costs. This professional standard has its own scope of applicability. The economic point is that each cost category may respond to inflation through a different driver and time horizon.

A simple net underwriting illustration defines fixed-dollar expected cost F as retained claims and expenses independent of premium, plus the net external cost of reinsurance premium under a consistent convention. Let v be expenses proportional to gross premium P and q a target contribution margin as a fraction of gross premium. Assume the listed cost categories are mutually exclusive, and ignore investment income, taxes, timing and capital charges beyond the illustrative margin. Required premium is P = F/(1 − v − q), provided the denominator is positive.

In the baseline illustration, expected retained claims are $6 million, fixed expenses $1.2 million, and reinsurance premium cost $0.3 million. Variable expenses are 20% of gross premium, and the target contribution margin is 5%. The required premium is $7.5 million divided by 0.75, or $10 million. The ratio of total modeled cost to gross premium is 95%.

Assume prospective retained per-claim severity rises 8% and relevant frequency rises 2%, with fixed exposure. Expected retained claims become $6.6096 million. Fixed expenses rise 6% to $1.272 million, and the external reinsurance premium cost rises 10% to $0.33 million. The combined fixed-dollar cost is $8.2116 million, and required premium is $10.9488 million. This is a 9.488% increase over baseline.

If only $10.5 million premium is achieved, the cost-to-premium ratio becomes ($8.2116 million + $2.1 million)/$10.5 million = 98.2057%, leaving an illustrative contribution margin near 1.7943%. The cost ratio is defined for this scenario and is not a statutory or published combined ratio. In particular, statutory gross and net combined ratio conventions may treat reinsurance and premium denominators differently. The illustration uses retained losses plus the cost of purchased protection consistently and does not also charge gross losses.

This arithmetic distinguishes inflation in expected claims from the required premium movement and exposes the sensitivity of the margin. A rate change equal to the headline inflation rate may be excessive for some segments and insufficient for others. The relevant objective is the forecast cost and economic contribution of the specific underwriting decision.

Payment timing and duration

Claims with lengthy settlement horizons remain exposed to changes in future price levels. The magnitude of the response depends on the timing and inflation sensitivity of each payment. For an illustrative ground-up nominal base cost of one dollar measured at current prices, suppose 50% is paid in one year, 30% in two years, and 20% in four years. All amounts are assumed to scale with a common annual cost growth rate r until payment, with no change in payment shares or contract terms.

The prospective nominal multiplier is 0.50(1 + r) + 0.30(1 + r)^2 + 0.20(1 + r)^4. At 4% annual growth, it is 1.078451712. At 9%, it is 1.183746322. The second set of assumptions therefore increases the nominal total by approximately 9.76% relative to the first. The example isolates inflation timing; it does not represent the average experience of any actual line.

An economically realistic cash-flow forecast can split the claim into payment-specific cost categories. Some amounts may have already been negotiated or legally fixed, while others may depend on future wages or service prices. Outstanding claims may also respond to settlement acceleration, delay or changing legal practice. Applying a common growth rate until payment to every amount would be inappropriate where the amount becomes fixed earlier.

For valuation purposes, nominal future cash flows and nominal discount rates must be used consistently. A rise in discount rates can reduce present values and partially offset an increase in nominal claims liabilities, but it does not erase nominal cash outflows. EU Solvency II Article 77 specifies a probability-weighted discounted best estimate and separate calculation of reinsurance recoverables within that prudential framework. Its valuation requirements are distinct from prospective underwriting price and commercial liquidity. The provision as inspected supports this methodological separation; this research does not survey all amendments, transition provisions or national implementation.

Creedon et al. (2026) develop practical scenarios in their volatile inflation reserving guide. Its separation of historical estimation and prospective reserve assumptions reinforces the need for coordination without conflating underwriting and reserving. The underwriter's future new-business cost should incorporate expected accident dates and settlement paths; the reserve analyst's relevant population is outstanding obligations and their future payment timing.

Common shocks and capital sensitivity

An inflation shock can change the dependence structure within an insurance portfolio. Let Z represent a common future cost scenario, and let S denote aggregate loss. The law of total variance states that Var(S) = E[Var(S | Z)] + Var(E[S | Z]). The first component describes expected variation conditional on the scenario; the second describes variation in the scenario-dependent aggregate mean.

If n similar exposures are conditionally independent given Z, their conditional idiosyncratic variance grows approximately in proportion to n under the simplified identical-exposure case. The variation of the conditional aggregate mean can grow in proportion to n squared when its per-exposure mean is a common nonconstant function of Z. Hence diversification of physical event risk does not necessarily diversify the systematic cost shock.

This result is an application of standard probability, not a new capital model. Its underwriting significance is that geography or product diversification must be examined relative to the cost-disruption mechanism. Multiple property portfolios may use common construction supply chains; liability portfolios may share a legal environment; motor and property may be jointly exposed to repair labor constraints. Conversely, lines using different inputs and currencies may respond differently. Dependence assumptions require evidence and stress testing.

The IAIS (2024) enterprise risk management principles place underwriting policy within risk management for solvency purposes and include attention to claims and expense behavior, macroeconomic conditions, limits and reinsurance. The framework is international supervisory guidance; its force in a given jurisdiction depends on implementation. It supports assessing common cost shocks through the full portfolio rather than considering each underwriter's book in isolation.

Capital and liquidity effects also depend on assets. When inflation coincides with higher interest rates, fixed-income asset values may fall while liability present values fall or rise depending on duration, discounting and cash-flow changes. Short-term liquidity may deteriorate despite a favorable accounting or solvency valuation effect if claim payments increase, collateral is required, or asset sales realize losses. These mechanisms belong in integrated financial analysis.

Scenario design and prudential oversight

The PRA's 2024 2024/25 business plan described supervisory follow-up on claims inflation and the risk of optimistic assumptions in pricing, reserves and capital. Its relevance is historical evidence of supervisory attention in a defined market. A forecast can be internally coherent yet optimistic if it assumes rapid mean reversion, perfect premium pass-through, unchanged reinsurance availability and no adverse change in selection simultaneously.

A baseline scenario should be separated from a stress scenario and from reverse stress testing. The baseline represents the institution's central assumptions and associated uncertainty. A stress specifies an adverse but interpretable path for key drivers. Reverse stress explores what combination of events would cause an identified financial or operational constraint to be breached.

An underwriting stress test may include a price-level jump in parts, persistent wage growth, a medical utilization change, delayed repricing, and increased reinsurance cost. Its components must be compatible. The analyst should map assumptions to cash flows and contracts, avoiding overlapping charges such as applying both a full severity trend and the same component-specific index. An apparently severe scenario can be economically inconsistent if one assumes that exposure counts remain fixed despite large changes in price and eligibility, unless this is intentionally stated as a conditional stress.

The result of the stress test should be expressed in quantities relevant to decisions: expected and tail losses, risk-adjusted contribution, exposure concentration, capital or liquidity needs, contractual protection, and timing. A change in the mean alone provides limited information about upper layers, aggregate exhaustion and recovery uncertainty.

Comparative synthesis, limitations and reproducibility

International comparability and institutional context

International market comparisons provide evidence about the coexistence of rising claim costs and underwriting responses, but they are shaped by product definitions, insurance penetration, inflation measures, valuation techniques, accounting and regulatory rules. OECD's (2025) non-life underwriting performance review reports jurisdiction-specific developments for 2024, including differences in the role of premium changes, claims experience and reinsurance costs. The report's comparative observations support examining the interaction of these mechanisms; they do not establish causal pass-through coefficients or a globally uniform timing.

A comparable study across markets must first fix the question: whether to examine cost per unit of covered exposure, rates per standardized risk, written premiums, earned premiums, loss ratios, combined ratios or retained risk. Local currency growth can differ from common currency growth. Gross observations may differ from net observations. Aggregated insurers may include different business mixes. Inflation benchmarks may use different statistical scopes or methods.

The United States, United Kingdom and European Union illustrate different institutional functions in the sources used here: US professional actuarial standards and statistical measurement; UK market and conduct oversight; EU prudential valuation and aggregate evidence. IAIS principles connect this comparison to a broader international framework. The sources complement each other without being interchangeable. The study does not establish a uniform global law of renewal or valuation and does not claim that these examples represent every insurance market.

Analytical synthesis of transmission mechanisms

The original analytical contribution of this study is the systematic connection between cost measurement and actual underwriting decisions. Price changes originate at the level of economic resources. They influence the amounts and quantities involved in covered events. Contractual transformations determine the insurer's obligation. Portfolio composition determines the relative frequency of different obligation types. The settlement horizon determines when cost changes enter cash flows. Risk transfer determines the allocation between insurer and reinsurer. Premium earning and regulation determine how quickly additional revenue can be generated.

Each of these mechanisms can be independently investigated within established actuarial and economic methods. Their combined effects explain why underwriting profitability cannot be inferred from an aggregate inflation index or an aggregate claims-cost series. This synthesis does not institute a new theory or claim priority. It translates existing knowledge into a coherent analysis of prospective contractual and financial decisions.

Several practical distinctions follow. A high nominal claim trend with stable fixed-mix intensity can indicate changing composition rather than pure price pressure. An inflation-indexed insured value can raise coverage adequacy while leaving significant retained loss sensitivity. A premium increase can improve expected revenue while the earning lag and selection effects reduce its first-year contribution. A favorable discounted liability valuation can coexist with deteriorating prospective underwriting costs. These results arise from different dimensions of the same decision and are compatible.

Measurement design for empirical application

A future insurer-level study could start with a versioned panel of risks, policies, claims and payments that share consistent identifiers. The policy record would contain inception, expiry, coverage, deductibles, limits, exposure bases, location and classification. Claims records would retain event, report, valuation and settlement dates, amounts paid, case estimates and recoveries. Procurement observations would distinguish quantity and price when available. The analysis would operate at a defined valuation date to prevent future information from entering earlier decisions.

To estimate severity trends, the analyst could construct cohorts by class, coverage, accident period and claim maturity. Changes in the distribution of covariates should be documented, with separate examination of fixed-mix and current-mix measures. When thresholds change, the truncation and censoring mechanism should be represented explicitly; where unobserved ground-up losses remain unidentified, the evidence boundary should be reported.

A conditional model can relate relevant cost indexes and operational variables to severity after adjusting for observed composition. Index lags should reflect the economic process and be selected using information available at the estimation date. Forecasting performance should be assessed across temporal holdouts, not random splits that mix future and past observations. Prediction intervals and scenario ranges should reflect estimation error, index uncertainty and the possibility of regime change.

The ASB's (2016) data quality standard addresses data selection, review and disclosure in applicable professional work. Its (2019) modeling standard addresses model purpose, assumptions, validation and communication. These support documenting whether a model is suitable for the actual underwriting decision. They do not confer validation on an unestimated scenario illustration.

A credible empirical evaluation should also examine contract terms: whether shifts in limit distributions, deductibles, endorsements or settlement practices can explain an apparent change in measured severity. Claims and exposure data should be reconciled with pricing and financial aggregates using consistent denominators. Differences that cannot be reconciled should be retained as explicit uncertainties, rather than absorbed into an apparently precise inflation residual.

Reproducible scenarios and analytical checks

All numerical examples use invented inputs and deterministic arithmetic. They can be independently reproduced by the following procedure.

  • Frequency–severity: multiply the two growth factors, subtract one.
  • Cost basket: multiply each baseline component share by its respective percentage price increase and add the contributions.
  • Deductible and cap: for each ground-up loss, subtract the deductible, floor the result at zero, cap it at the maximum insurer payment, multiply by probability, and add.
  • Mix: calculate baseline mean under baseline weights, new mean under baseline weights, and new mean under new weights. Subtract consecutive results.
  • Reinsurance: transform each of the three event losses through the stated excess layer and subtract the total ceded payout from total gross loss.
  • Premium adequacy: forecast each mutually exclusive retained or external cost component and divide their sum by one minus the variable expense ratio and target margin.
  • Duration: compound each payment share over its stated horizon and add.
  • Change all assumptions one at a time as a diagnostic, then apply coherent multivariate scenarios to retain interactions.

The examples use exact arithmetic until results are displayed; percentages are rounded only for readability. The assumed coverage conventions are stated rather than inferred from market practice. The sensitivity identities depend on the stated conditions; independent conditional exposures in the variance illustration are a mathematical construction, not an empirically established dependence structure.

Evidence quality and limitations

The primary evidence sources have distinct scopes. Professional standards describe applicable practice requirements. Supervisory material describes institutional expectations or specific rules. Official statistics identify the quantities actually measured. Scholarly papers investigate claims inflation using defined methods and settings. Industry studies supply market-specific estimates and interpretations. Documentary evidence establishes the plausibility and importance of mechanisms but cannot estimate their joint magnitude for an unspecified insurer.

The research does not possess a global harmonized microdata set, insurer-specific policy and loss histories, proprietary claims development triangles, or confidential reinsurance terms. It therefore does not fit inflation coefficients, estimate confidence intervals, calculate an actual insurer's capital requirement, identify the causal effect of social inflation, or establish the financial value of a particular renewal strategy.

Bohnert et al.'s peer-reviewed record is supported by the separately identified publicly accessible prepublication manuscript rather than an audit of the final paywalled text. Taylor's research is cited at the level of its accessible publisher extract. Industry social inflation sources are cited for their explicitly reported estimates and subjects, not independently reproduced. Official statistical and supervisory sources are interpreted only for the sections and jurisdictional scope identified in the evidence ledger.

External expert review and field validation are outside the scope of this authorial research. The limitations restrict quantitative claims and prescription. They preserve the usefulness of the general economic and actuarial synthesis by establishing precisely what the available evidence supports.

Operational implications

Underwriting sensitivity analysis should be incorporated into the ordinary governance of risk acceptance and renewal. A portfolio can establish the material cost drivers of its dominant insured losses, monitor the evidence for changes, create consistent prospective scenarios, and evaluate their consequences through policy and reinsurance terms. The scope of each forecast should match the decision horizon.

The most valuable analytical output is often a range of feasible outcomes rather than a single expected trend. The range should distinguish uncertainty about economic prices, uncertainty about pass-through, uncertainty about contract-level loss distributions and uncertainty about revenue response. A decision that remains adequate under plausible alternative assumptions has a different reliability profile from a decision dependent on a narrow central scenario.

Responsibilities can be delineated without isolating departments. Claims professionals hold operational information about settlement and procurement. Actuaries integrate loss data and prospective trends. Underwriters control acceptance, terms and execution within authority. Risk and finance teams evaluate capital, reinsurance and financial resilience. Their assumptions should be reconciled in a traceable process.

Material triggers for reviewing underwriting decisions include evidence that cost indexes no longer represent insured inputs, a change in the distribution of positive-paid claims, rising frequency of limit exhaustion, a divergence between technical and achieved rate, and concentrated sensitivity to a common economic shock. A trigger initiates investigation; it does not dictate a particular price or coverage action. Such action depends on evidence, authority, law and the rights of the affected parties.

Conclusion

Inflation changes the prospective economics of insurance by altering the resources required to fulfill insured promises, their payment timing and their distribution across contracts. Underwriting sensitivity is consequently a property of a specified portfolio, contract structure, forecast horizon and institutional environment. A headline price measure becomes decision-relevant only through a defensible mapping to the costs and obligations it represents.

The first research question concerns pass-through and claim composition. The analysis establishes that ground-up price movements, quantities used, event frequency, changing selection, and contractual thresholds can jointly alter expected claims and tail risk. Fixed deductibles can amplify percentage changes in positive payouts; payment caps can attenuate them; excess layers can respond in different proportions from gross loss. These effects are shown using deterministic calculations rather than asserted as empirical estimates.

The second question concerns index selection. Relevant benchmarks correspond to the measured inputs, market, geography, currency and time horizon of covered claims. Repair materials, compensation, medical services, legal expenses and supply-chain disruption require different evidence. A spending series, an input price measure and a residual claims trend describe different quantities. Their distinctions should remain visible from the original source through the prospective cost forecast.

The third question concerns the timing and nature of contractual adjustment. Renewal terms should be reassessed when evidence indicates a material change in prospective risk costs or coverage adequacy, and changes must be evaluated alongside earning lags, selection, available treaty protection, legal constraints and policyholder outcomes. The illustrative premium analysis demonstrates the effect of a modest shortfall in achieved revenue on the contribution margin. The duration and variance analyses show why future costs and common scenarios retain significance beyond the first annual pricing period.

The fundamental conclusion is that inflation adequacy is an underwriting and portfolio problem involving the alignment of cost expectations, contract obligations, exposure measurement, revenue realization and financial risk-bearing capacity. This alignment is maintained through evidence, prospective analysis and ongoing reassessment. The monograph's contribution is the integrated interpretation of these established mechanisms within the specific scientific object of underwriting sensitivity. Its quantitative examples are reproducible, its documentary conclusions remain bounded by the sources examined, and its findings are available for future empirical testing with appropriately governed insurance data.

References

  • Actuarial Standards Board. (2009). Actuarial Standard of Practice No. 13: Trending Procedures in Property/Casualty Insurance (June 2009). Source record

  • Actuarial Standards Board. (2016). Actuarial Standard of Practice No. 23: Data Quality. Source record

  • Actuarial Standards Board. (2017). Actuarial Standard of Practice No. 53: Estimating Future Costs for Prospective Property/Casualty Risk Transfer and Risk Retention (December 2017). Source record

  • Actuarial Standards Board. (2019). Actuarial Standard of Practice No. 56: Modeling (effective 2020). Source record

  • Benigno, G., di Giovanni, J., Groen, J. J. J., & Noble, A. I. (2022). The GSCPI: A New Barometer of Global Supply Chain Pressures. Federal Reserve Bank of New York Staff Reports, No. 1017. Source record

  • Bohnert, A., Gatzert, N., & Kolb, A. (2015). Assessing Inflation Risk in Non-Life Insurance. Prepublication manuscript, October 24, 2015. Source record

  • Bohnert, A., Gatzert, N., & Kolb, A. (2016). Assessing Inflation Risk in Non-Life Insurance. Insurance: Mathematics and Economics, 66, 86–96. doi:10.1016/j.insmatheco.2015.11.003. Source record

  • Centers for Medicare & Medicaid Services. (2026). National Health Expenditure Data: Historical, 2024 Estimates (January 14, 2026 release). Source record

  • Creedon, C., Ahuja, K., Bargate, E., Lenney, S., Orenstein, E., Schofield, M., & Stock, R. (2026). Reserving for volatile inflation: a practitioner’s guide to setting best estimate reserves during periods of inflation volatility. British Actuarial Journal, 31, e18. doi:10.1017/S1357321726100518. Source record

  • Creedon, C., Bargate, E., Lenney, S., Schofield, M., & Stock, R. (2025). Claims inflation estimation: a practical guide for historical data. British Actuarial Journal, 30, e31. doi:10.1017/S1357321725100275. Source record

  • European Insurance and Occupational Pensions Authority. (2023). Impact of Inflation on the Insurance Sector. EIOPA-BoS-23-360, October 5, 2023. Source record

  • European Union. (2009). Directive 2009/138/EC (Solvency II), Article 77: Calculation of Technical Provisions. EIOPA Single Rulebook. Source record

  • Financial Conduct Authority. (2022). Insurance: Conduct of Business Sourcebook, ICOBS 6B.2, Setting Renewal Prices, rule 6B.2.1R (effective January 1, 2022). Source record

  • International Association of Insurance Supervisors. (2024). Insurance Core Principles and ComFrame: ICP 16, Enterprise Risk Management for Solvency Purposes. Source record

  • Lloyd’s. (n.d.). Principle 1: Underwriting Profitability. Source record

  • Lynch, J., & Moore, D. (2023). Social Inflation and Loss Development—An Update. Casualty Actuarial Society and Insurance Information Institute, March 6, 2023. Source record

  • National Association of Insurance Commissioners. (n.d.). Model Laws, Regulations and Guidelines. Source record

  • OECD. (2025). Global Insurance Market Trends 2025: Underwriting Performance in the Non-Life Insurance Sector. OECD Publishing. Source record

  • Prudential Regulation Authority. (2024). PRA Business Plan 2024/25 (April 2024). Bank of England. Source record

  • Swiss Re Institute. (2024). Social Inflation: Litigation Costs Drive Claims Inflation. sigma No. 4/2024. Source record

  • Taylor, G. C. (1977). Separation of Inflation and Other Effects from the Distribution of Non-Life Insurance Claim Delays. ASTIN Bulletin, 9(1–2), 219–230. doi:10.1017/S0515036100011533. Source record

  • U.S. Bureau of Labor Statistics. (2022). Inputs to Industry Price Indexes (June 21, 2022). Source record

  • U.S. Bureau of Labor Statistics. (2025). Handbook of Methods: Employment Cost Index, Overview (September 30, 2025). Source record

  • U.S. Bureau of Labor Statistics. (n.d.). Measuring Price Change in the CPI: Medical Care. Source record