Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Catastrophe Bond Secondary Market Liquidity

Abstract. Secondary-market liquidity determines the price, time, and execution conditions under which holders of outstanding catastrophe bonds can transfer catastrophe-linked exposure and obtain usable cash. It influences yield spreads, portfolio construction, financial stability, investment-fund redemptions, and the allocation of catastrophe-risk capital. This monograph investigates the international catastrophe-bond market through contractual analysis, financial-market microstructure, transaction-based and proxy-based academic studies, institutional source records, market reporting, and the European discussion of UCITS eligibility. It distinguishes trading activity, quoted spread, depth, immediacy, resilience, liquidity premium, model valuation, and redemption liquidity. Herrmann and Hibbeln (2023) estimate a liquidity component of approximately 21 percent of observed yield spreads in their TRACE-based sample; Zhao and Yu (2019) estimate 9.42 percent using indirect proxies and a different historical dataset. These findings establish meaningful liquidity compensation under their respective designs without supplying a universal coefficient for present-day transactions. Aon, Swiss Re, EFAMA, and ESMA provide evidence of market expansion, observed transaction activity, and persistent concerns about stress execution. Hypothetical, reproducible examples distinguish losses caused by changing catastrophe expectations from concessions needed for immediate liquidation and demonstrate the cash-flow consequences of finite market depth. The analytical conclusion is that catastrophe-bond liquidity is a contract-specific, quantity-dependent, time-dependent, and event-conditional execution capability. Its rigorous measurement requires simultaneous attention to security terms, fundamental risk information, the investor base, actionable quotes, transaction data, and the liquidity obligations of investment vehicles. No new proprietary theory, original empirical dataset, or unverified causal estimate is asserted.

Keywords: Catastrophe bonds; insurance-linked securities; secondary-market liquidity; trade execution; bid–ask spread; liquidity premium; catastrophe risk; market making; TRACE; OTC markets; fund redemptions; valuation; UCITS; price impact; event uncertainty; market depth.

Introduction

Catastrophe bonds permit investors to bear defined catastrophe exposure in exchange for contractual compensation, thereby connecting insurance risk transfer with capital markets. A sponsor obtains protection against a specified peril or class of losses; investors acquire notes that may experience principal reduction if the contractual event trigger is reached. The NAIC insurance-linked securities overview describes the broader institutional family, while the World Bank capital-at-risk notes program describes a sovereign and multilateral application. The value of the note changes as the remaining exposure, event information, contractual time limits, discount rates, and demand for catastrophe risk evolve.

The secondary market addresses the transfer of a position after issuance. It is distinct from new placement, because a secondary purchase ordinarily reallocates an existing contractual risk-bearing position among investors without expanding the sponsor's already arranged protection. This distinction separates primary capital formation from the efficiency with which outstanding risk capital can circulate. The scientific question is therefore what enables a specific holder to sell a specific catastrophe bond for a specified quantity, by a specified deadline, at an economically understood price.

A large asset class may contain many thinly traded individual notes. An investor may be willing to hold to maturity, find no reason to sell, and yet receive attractive bids if it decides to exit. Conversely, a security with many historical trades may become costly to sell when several large holders simultaneously seek cash. Volume, turnover, price marks, and executable depth therefore occupy different evidentiary levels. Understanding their relationship is necessary to evaluate reported market liquidity.

Froot's (2001) examination of catastrophe-risk capital relates catastrophic loss financing to constraints on risk-bearing supply, illuminating the capital allocation underlying the catastrophe-bond market. Braun (2016) studies primary cat-bond spread formation, including expected-loss and institutional variables. These analyses help define the initial equilibrium compensation for taking a catastrophe risk. Once the security trades, exit costs, investor constraints, and evolving event information also matter.

The strongest direct study in the source record is Herrmann and Hibbeln (2023), whose transaction-level liquidity research uses FINRA TRACE observations for selected U.S.-reportable catastrophe bonds. Zhao and Yu (2019), in the liquidity-premium proxy study, infer compensation from secondary spread data and indirect measures. Their empirical results address related but nonidentical populations. Treating them as interchangeable would mistake methodological variation for historical change.

Contemporary industry reporting documents how the market's scale has developed beyond those research samples. Aon's August 2026 market release reports approximately $63.4 billion outstanding as of June 30, 2026, following $24.9 billion of issuance over the preceding twelve months. The Artemis catastrophe-bond reports chronicle a still-changing issuance market through Q3 2026. Outstanding amount and new issuance confirm risk-capital scale; neither establishes the ability to dispose of any particular block at a stable price under stress.

This study follows the six approved research chapters. It identifies contractual and market units, analyzes quotations and cash flows, examines information following disaster events, assesses intermediaries and empirical premiums, investigates open-ended fund obligations, and integrates the institutional debate with the research record. Mathematical calculations are explicitly hypothetical. Publisher records and policy reports are attributed to their producing institutions; recommendations are distinguished from binding law. The evidence cutoff is November 14, 2026.

Definitions, Source Record, and Scientific Boundary

Contractual security and transferred exposure

A catastrophe bond is a debt-like security whose coupons and principal depend on specified insurance-related events. Its transaction documents may designate an indemnity trigger, an industry-loss index, physical parameters, modeled losses, or combined conditions. The trigger and attachment terms determine when the holder may lose principal and when the sponsor receives protection. Collateral arrangements, eligible investments, return calculation, legal vehicle, maturity, reporting rights, and event determination provisions differ across deals. The 2026 Jamaica catastrophe-bond placement illustrates a specific World Bank transaction with explicitly described parametric coverage; its legal terms cannot be presumed to govern indemnity or industry-loss bonds.

The secondary traded object consists of the rights and obligations of a particular outstanding security at its current state. Its original $100 face value may no longer represent $100 of current principal outstanding if an event has already caused a contractual reduction. A quoted price must be identified by the principal base and settlement convention to which it applies. Remaining event exposure, coupon accrual, principal impairment, and the resolution of an event notice are economically relevant attributes of the object being sold.

An investor's sale generally transfers an investment interest while keeping the originating sponsor's protection within the established security structure. The arrangement remains subject to contractual transfer restrictions, custody and settlement rules, and applicable market and regulatory frameworks. The IBRD capital-at-risk note structure illustrates why the same underlying catastrophe financing can have a sponsor-facing protection purpose and an investor-facing tradable note. Whether the sponsor can alter protection following a transfer depends on the particular documentation rather than a general secondary-liquidity rule.

A precise vocabulary of liquidity

Immediacy refers to how quickly a trade of a defined size can execute; depth refers to quantities available at stated prices; tightness refers to the price difference between an executable bid and offer; resilience refers to recovery of market conditions after a trade or shock; and breadth concerns the range of participants and securities. These categories describe different features and may diverge. A portfolio holding can have a narrow quoted bid–ask spread but little depth, while another has a wider spread and deeper firm bids.

A quote is actionable only to the extent that a counterparty commits to the specified size and terms. Indications, dealer valuation sheets, broker runs, evaluated prices, and executable bids are all useful information, yet they have different evidence content. A midprice calculated from indicative bids and offers does not guarantee that a seller can achieve the bid for the portfolio's entire position. Price transparency and transaction capacity are related but distinct properties.

Turnover is a ratio of properly counted traded notional to comparable outstanding notional over a defined period. It measures activity conditional on the reporting universe, not the demand curve a prospective seller would face. Zero trades can mean no buyers or no willing sellers. Conversely, frequent small trades do not establish block-trade capacity. A careful empirical paper reports units of trade size, the time window, the eligible security population, and treatment of capped transaction sizes.

Liquidity premium refers to the component of required return attributed to investors' expectations of trading costs or liquidity conditions. Yield risk spread refers to compensation beyond a benchmark yield and may include expected catastrophe loss and catastrophe risk premium. Bid–ask spread refers to an immediate pricing wedge. The three cannot be treated as numerical synonyms. The conceptual relation between required returns and liquidity risk in asset pricing is well established, but the magnitude of that relation must be estimated afresh for a catastrophe-bond population.

Source classes and reproducibility

This study distinguishes five kinds of source. Contract and institutional records establish the economic instrument; statutes and supervisory documents establish legal obligations or regulatory proposals; peer-reviewed empirical studies support findings within their samples; intermediary reports provide industry observations, often compiled from proprietary data; and investment-industry policy papers articulate evidence and advocacy that require source attribution. The different sources are compared without assuming that all are independent or equally complete.

FINRA's TRACE reporting documentation establishes the infrastructure for eligible over-the-counter fixed-income trade reporting. It does not imply that every catastrophe-bond transfer, every failed trade, or every indicative quote is represented. Herrmann and Hibbeln's use of TRACE requires careful matching of identifiers and dealer transactions; their selected observations are informative about observed executions and do not by themselves reveal the global unexecuted order flow.

For market-microstructure analogies, Amihud (2002) proposes a price-impact measure based on volume, while Acharya and Pedersen (2005) investigate how exposure to changing trading conditions can enter asset prices. These foundations motivate research questions, rather than supply ready-made catastrophe-bond effect sizes. Rare trading, peril-specific news, risk-model revisions, and distinct investor mandates must be included explicitly when adapting conventional securities methods.

The study adopts a documentary and critical-comparative method, a formal cash-flow method for illustrative examples, and an evidence hierarchy for empirical claims. It does not reconstruct privately inaccessible offering memoranda or attempt statistical replication without an authorized transaction dataset. Claims based on publisher abstracts or industry releases remain identified by that source level. Sources after the research cutoff and facts without a checked research record are excluded.

Scientific object and related research boundaries

The object comprises tradeability and pricing of outstanding catastrophe bonds, market intermediation, the influence of event information, the relationship between trading costs and yields, and investors' need to raise cash through secondary sales. Primary issuance, catastrophe-loss modeling, insurance claims settlement, collateral trapping, fund law, and broader insurance finance enter the analysis when their mechanisms affect secondary liquidity. Each remains an independent field of research. No novel system of catastrophe-bond liquidity is postulated; the monograph integrates existing contractual, empirical, and institutional knowledge.

Trade Conventions and Secondary Prices

Secondary dealing and execution

In specialist over-the-counter dealing, market participants identify a note, desired side, amount, and execution window, and obtain indications or bids from dealers and brokers. A matching intermediary searches for another investor with complementary interest. An inventory-taking intermediary buys or sells principal and may hold the exposure before offsetting it. Both forms can facilitate liquidity, but they allocate risk and cost differently.

Specialized buyer mandates restrict possible counterparties. Peril concentration, trigger type, sponsor exposure, issue size, risk-model familiarity, seasonality, currency, remaining maturity, and expected loss can all determine whether a mandate permits a purchase. The economically relevant measure of market depth is therefore buyers able and willing to absorb the particular marginal exposure at the proposed terms, not the total value of all outstanding catastrophe bonds.

Aon's 2025 secondary-market report describes strong investor demand and relatively limited offered securities during parts of the first half of 2025. It estimates about $2–$3 billion of secondary trading in that half-year, excluding interfund cross trades. This is a proprietary market description under the provider's chosen definitions, not a comprehensive FINRA count. Strong bidding can coexist with limited trading when existing holders prefer to retain their positions.

Transaction price and contractual notional

A clean quote and a dirty settlement cash price differ by accrued income and deal-specific adjustments. In a simplified calculation, with current relevant principal N, clean price P per $100, and accrued coupon A, gross cash consideration is N(P/100)+A. The actual transfer consideration must also follow the note's definitions, settlement date, reductions of principal, collateral income, and relevant fees. A bid relative to reduced principal cannot safely be compared with a quoted original-face-value price without normalization.

For a schematic single-period example, let the catastrophe event probability be p, principal loss fraction conditional on the event be λ, and current $100 principal be payable at the terminal date. The expected principal repayment before discounting is 100(1−pλ). This identity is valid for the stated two-state construction; real cat bonds can have multiple covered events, per-occurrence or aggregate features, an industry-loss index, attachment and exhaustion points, or event development processes. The expected-loss variable must be produced by a documented model and contractual definition for any actual note.

The present value is the conditional discounted expectation of all remaining contractual coupons and principal under assumptions about reference rates, conditional losses, and investors' required compensation. The price can therefore change because modeled loss risk changed, market risk compensation changed, rates changed, or execution liquidity changed. A residual difference between a quoted bid and a model mid is not automatically a pure liquidity effect: the model itself can be stale, incomplete, or calibrated to market liquidity.

Braun's primary-market spread evidence reports expected loss as a principal driver of initial cat-bond spreads, with territory, sponsor, cycle, and broader financial variables also relevant. Secondary traders reprice the surviving risk over the shorter remaining life, incorporating newly available meteorological, geological, and insured-exposure information. Differentiating this repricing from an exit concession is the central identification difficulty.

Risk spread, bid–ask spread, and price sensitivity

Suppose a floating-rate catastrophe note pays a benchmark investment return plus a risk spread. A widening of required risk spread generally lowers the price of an otherwise unchanged note, but the correct valuation requires discounted event-dependent cash flows. In the elementary price-duration approximation ΔP/P ≈ −D_mod Δy, the modified-duration factor depends on remaining cash flows and discount assumptions. A $100 note with hypothetical effective spread duration of 2.3 years experiences approximately a 2.76 percent price reduction when the relevant required spread widens by 120 basis points, all else held equal. The example is not a market observation or model-calibrated catastrophe-bond duration.

The instantaneous dealer bid–ask price wedge is a second dimension. Hypothetical quotes of 99.40 bid and 100.20 offer per $100 of current outstanding principal have a mid of 99.80. The 0.80-point wedge equals approximately 0.802 percent of the mid. A seller transacting $5 million current principal at 99.40 would receive $4.97 million before accrued amounts and fees; a buyer paying 100.20 would pay $5.01 million on that price basis. The $40,000 comparison illustrates the contemporaneous spread at the same size, not an empirically verified fee or assured sequential round-trip cost.

Crucially, quoted prices must be accompanied by executable quantity. If 99.40 bid applies only to $500,000 current principal, the $5 million proceeds in the preceding calculation are a hypothetical extrapolation. A dealer willing to acquire the next $4.5 million may require a larger concession. Price schedules by quantity, rather than a single top-of-book mark, determine actual block liquidation cost.

Consider a separate hypothetical comparison. Security A has a bid–ask price spread of 0.40 points and firm bid depth of $100,000, while Security B has a 0.90-point spread and $10 million of firm bid depth. A holder selling $10 million cannot infer that A is cheaper to liquidate. A has a tighter first-lot quote; B may have greater depth for that holder's proposed trade. The instrument and trade-size dependence of measured liquidity should remain explicit in any cross-security ranking.

Price sources, time stamps, and stale marks

The relevant observations may be genuine executions, dealer bids, nonbinding broker runs, composite price sheets, or a model valuation. A secondary price index may report smooth marks through a quiet period because few trades challenge its inputs. After a major event, discontinuous repricing may reveal both catastrophe loss information and previously unseen liquidity costs. As Aon Securities reported in 2025, actual bids in a strongly demanded market could exceed some generic sheet marks. Such evidence demonstrates that a valuation sheet and the best executable bid need not coincide in either direction.

A research dataset should preserve instrument identifier; original and current principal; issue and maturity dates; trigger type; risk period; price convention; timestamp; dealer versus customer side; size; currency; execution status; accrued amounts; available quotes; known catastrophe-event exposures; and subsequent loss updates. Absent fields constrain interpretation. A note-month observation, a completed dealer round trip, and an individual trade must never be collapsed into one statistical unit without an explicit method.

Liquidity under Event Uncertainty

Disaster information and conditional value

The physical events protected by catastrophe bonds can change expected principal repayment rapidly. Hurricane track forecasts, wind fields, earthquake intensity measurements, catastrophe model revisions, insured-exposure uncertainty, claims development, and index-loss updates do not arrive simultaneously. Some information precedes landfall, while final insured losses may be revised for months or longer. An investor valuing an outstanding note must specify the information available at the trade time, not use a later final loss estimate as though it were known to participants.

A price decline after an event may be a fundamental revaluation of expected contractual cash flows. A separate concession may arise from disagreement among potential buyers, greater risk aversion, limited balance sheets, poor access to current exposure data, or sudden redemption demands. Both effects can occur at once. Gürtler, Hibbeln, and Winkelvos document financial-crisis and catastrophe repricing in secondary CAT-bond spreads and show why an absolute claim of market independence from general financial risk is too strong for the documented sample. Low dependence of catastrophic physical loss on ordinary corporate earnings does not remove all financing and market-liquidity linkages.

Herrmann and Hibbeln (2023) report lower transaction activity in peak U.S. hurricane season, increased trading nearer maturity, and low trading volumes during several severe hurricane episodes. Their observed event-window trading results establish that more intense attention to loss events need not produce more executed trades. Participants may defer sales and purchases while the distribution of eventual impairment remains disputed. That behavioral and informational mechanism is consistent with the evidence, though the data do not justify a universal assertion that every catastrophe causes liquidity to collapse.

The probability of a principal reduction depends on the terms and current state. A high-level hurricane can leave one bond unaffected and put another near exhaustion. Regional proximity alone is not an exact exposure measure. Analysts should separate insured attachment and exhaustion, event eligibility, trigger measurement, and modeled probability of recovery. A lower bid for a bond with materially increased expected impairment cannot automatically be labeled illiquidity.

An explicit, conditional decomposition

Let V0 be a dated, documented pre-event modeled value; V1 be a comparable modeled value after event information, using a specified discount and liquidity convention; and B1 be the executable post-event bid for a defined amount. The difference V0−B1 can be written as (V0−V1)+(V1−B1). This is an accounting identity, but the economic meaning of the two components depends entirely on how V1 is specified. If V1 embeds trading costs or is based on stale quotes, the residual cannot represent a uniquely identified liquidity premium.

Hypothetical example: an event changes the model value from 100 to 92 per $100 current principal, while a firm bid for a specifically defined block is 88. The total observed fall from the chosen starting benchmark is 12 points. Eight points are attributed to the model's changed value and four to the bid-versus-model residual. A revised catastrophe model, a different discount curve, or later event measurements would modify the allocation. The 8/4 division is illustrative arithmetic, not a historical estimate of any actual catastrophe-bond trade.

An event may complicate principal determination without preventing a legal transfer. An indemnity trigger can rely on claims development; an industry-loss trigger can depend on index publications and revisions; a parametric trigger relies on validated physical data; a modeled-loss trigger requires particular input and calculation conventions. Each can change information latency, but legal settlement, potential extensions, and collateral release follow actual terms. No general rule orders all trigger types from most to least liquid.

Seasonality, maturity, and funding

Risk-bearing capacity is seasonal because hurricane exposure and investor rebalancing needs vary through the year. The effect of maturity is conditional: as an uneventful risk period ends, the remaining uncertainty may shrink and a note may become easier to trade. But a bond near contractual maturity with an unresolved covered event may be hard to price and may not return principal on a simple calendar schedule. Average statistical effects near maturity should not substitute for contract-specific loss and redemption provisions.

Swiss Re's July 2026 ILS market update reports a strong issuance environment in a comparatively favorable first-half market. Aon's 2026 outstanding-capital statistics similarly describe expanding market capacity. Neither source tests the precise ability of all large investors to exit simultaneously following a sequence of severe events. Issuance capacity can be robust while stressed secondary execution is shallow. The issue is empirical and state-specific, not resolved by the gross size of the market.

Fund cash demand can amplify event-related uncertainty. A catastrophe may reduce a portfolio's modeled NAV; investors may request withdrawals; fund managers may then sell other bonds to meet payments; and redemptions can generate secondary-sale pressure even in notes that suffered no direct principal loss. Insurance claims payments and fund redemption payments are distinct liabilities attached to distinct legal entities. Trapped reinsurance collateral elsewhere in an ILS strategy may affect portfolio cash availability without being identical to the assets or legal mechanism of a particular catastrophe bond.

The broader concept of common liquidity exposure in Acharya and Pedersen's pricing study provides a plausible analytic question: do catastrophe bonds become harder to sell when investors most value immediate cash? Answering it quantitatively would require trading and cash-flow evidence across ordinary and stressed states, not the mere observation that some bonds have high coupons.

Research design for event studies

An event-window study should predefine the timing of a physical event and relevant forecasts, specify which notes are exposed under contract terms, track dated loss estimates, and select comparable control notes or markets with disclosed selection criteria. Outcomes must be measured separately: executed prices, probability that an order fills, fill time, quoted depth, bid–ask spread, and post-event price recovery. Necessary controls include outstanding notional, expected-loss estimate, rating if available, remaining risk period, peril type, event exposure, and general market risk appetite.

Special attention should be paid to selection into trading. Transactions reveal the prices of deals actually completed; they do not directly reveal offers that were refused or severely impaired notes for which no investor would supply a bid. Using only realized spreads may underestimate the cost of exiting a stressed portfolio. Conversely, interpreting all post-event price falls as liquidity shocks overstates illiquidity by treating fundamental expected loss as a trading friction. The methods and sample coverage must determine which claims an empirical result can support.

Market-Maker Inventory and Bid–Ask Spreads

Intermediation economics

An intermediary can match buyers and sellers as an agent, hold principal inventory for resale, or commit to a quote contingent on limits. Each approach has a different balance-sheet burden. An inventory-taking dealer may face catastrophe risk while warehousing a bond and can demand compensation for inventory volatility, financing, information asymmetry, and settlement uncertainty. A matching broker may bear less open catastrophe exposure but still incur search costs, counterparty risks, and operational expenses.

Herrmann and Hibbeln (2023) identify many matched dealer buy–sell pairs occurring within roughly an hour. Their TRACE dealer round-trip analysis suggests that a substantial fraction of observed intermediation resembles broker matching rather than prolonged principal inventory holding. The inference refers to a selected transaction sample and matching procedure. It does not establish that dealers universally abstain from carrying inventory or that every spread is a pure commission.

Broker competition also depends on the range of investors for a note. A high expected-loss bond concentrated in a narrow peril region may have fewer acceptable holders than a diversified note with an otherwise similar coupon. Dealers can match risk mandates only if capital is available and investment restrictions permit the purchase. Thus liquidity may vary by security characteristics even when average market turnover remains stable.

Spread and impact metrics

For a contemporaneous, executable bid B and offer A at a common size, the proportional quoted spread is (A−B)/((A+B)/2). The realized dealer spread requires paired transactions and an explicit attribution of sides, dates, principal base, and inventory holding interval. The effective spread compares an execution price with an appropriate contemporaneous mid, which can be uncertain for a thinly traded OTC security. These measures capture different frictions and must be reported under their precise definitions.

Turnover for period T is observed traded notional divided by correctly defined eligible outstanding notional for that period. Transaction-count frequency, normalized dollar volume, and volume-based price impact should not be substituted for this ratio. FINRA's trade-reporting documentation supports reconstructing the U.S. eligible reporting environment, but it does not guarantee visibility of every global catastrophe-bond transaction or unexecuted trade interest. If notional disclosure is censored or reporting is delayed, the resulting study must describe the bias and correction.

Amihud's absolute-return-to-volume measure is useful as a conceptual indicator of illiquidity. For catastrophe bonds, however, a large absolute return can reflect fundamental catastrophe news rather than mechanical order-induced price impact. Thin observations, composite marks, and zero-volume days can destabilize the ratio. Analysts should combine such proxies with actionably quoted size, actual executions, and event-adjusted expected-loss information, and they should avoid causal conclusions from the proxy alone.

Evidence of liquidity premiums

Zhao and Yu (2019), covering a historical 2002–2016 dataset, estimate an average liquidity premium of 67.57 basis points in a mean spread of 717.37 basis points, or about 9.42 percent, by using indirect measures such as time to maturity, volatility, and dispersion. Their liquidity premium estimates also associate market stress and severe catastrophes with liquidity compensation. The figures belong to the study's empirical model, variable definitions, and period; they should not be inserted unadjusted into contemporary investment valuations.

Herrmann and Hibbeln (2023), using direct transactions in their January 2015–March 2019 selected dataset, report 8,883 trades for 245 cat bonds and estimate liquidity compensation at approximately 21 percent of observed yield spreads, around 98 basis points on average and up to 141 basis points for the higher-risk group. Their transaction-based spread decomposition connects realized transaction frictions with observed spreads after documented controls. The selection of TRACE-eligible and identifiable notes affects external validity; different sample counts in preliminary working materials should not be confused with the published sample's numbers.

The 9.42 and 21 percent shares are not a clean time series. Their observation windows overlap, their empirical designs differ, and the relevant liquidity concepts are not exactly the same. One uses proxies with a broad historical dataset; the other analyzes realized trades in a narrower market. A direct claim that average liquidity compensation doubled in the intervening years would not follow from these papers. A scientifically defensible synthesis is that secondary liquidity can command a material yield premium, while its estimated magnitude depends on instruments, transaction data, modeled expected losses, and identification strategy.

Normal-state trading and order imbalance

A market can show low traded volume when willingness to hold is high and prospective buyers bid aggressively. In that case immediate liquidity for a modest seller can be strong even with few executions. A broad reversal of order flow changes the outcome: simultaneous sellers can face a steep downward demand curve if available investors are already concentrated in the same risks. Aon's 2025 secondary trading observations indicate that demand conditions must be considered alongside raw volumes. A single turnover rate is therefore an incomplete indicator of expected liquidation proceeds.

This market is also sensitive to the composition of supply. New catastrophe bond issuance competes with outstanding notes for investor capital, and maturity repayments create reinvestment capacity. When new issues offer improved risk compensation, a seller of an older note may have to offer a discount for its risk features. This discount can contain fundamental repricing and liquidity effects. To distinguish them, comparison notes must be matched for remaining risk period, expected loss, trigger, currency, and transaction size.

For any trader-facing assessment, the relevant outputs are a price-quantity curve, execution horizon, trading costs, and settlement reliability for the required side of the trade. Quoted market tightness at negligible size cannot guarantee efficient liquidation of a concentrated fund holding. Neither can aggregate ILS issuance growth guarantee persistent execution capacity under stress.

Investor Redemptions and Valuation Tests

Redemption liability and underlying bond markets

An open-ended fund may issue units redeemable under a timetable fixed by its documents and applicable law. The fund's obligation is to make the required payment on the relevant date according to those terms. A catastrophe bond in its portfolio may be tradable yet require substantial time or price concession to sell. Fund liquidity and note liquidity therefore must be assessed jointly but measured separately.

The consolidated UCITS Directive Article 84 addresses investors' ability to have units redeemed on request and provides for suspension in exceptional circumstances under applicable provisions. Its effects require analysis of local implementation, fund constitutive documents, and any later effective amendments. The legal framework does not assert that each constituent security is liquid at all times, nor does it specify an asset-level bid that is available to every manager.

ESMA's 2025 UCITS Eligible Assets advice discusses catastrophe bonds' specialized investor base, buy-and-hold characteristics, and exposure to event-related liquidity problems. This document is supervisory technical advice to the European Commission. It must not be described as a universally binding 2026 prohibition on catastrophe bonds in UCITS. Its underlying concern is that ordinary saleability may offer insufficient assurance about redemption obligations when large events or investor withdrawals coincide.

EFAMA's November 2026 UCITS market study makes a different, industry-association case. Drawing on market-participant estimates, it reports annual catastrophe-bond turnover of approximately 23 percent of outstanding, compared with around 33 percent for European investment-grade corporate bonds and 19 percent for European high-yield bonds. It also describes regular quotations, a specialist intermediary network, and established institutional investors. The comparison is descriptive: denominators, universes, trade-size composition, and observation procedures differ across asset classes. The reported turnover supports the existence of an active secondary market; it does not identify the worst-case executable discount for a particular redemption shock.

ESMA and EFAMA therefore address overlapping but distinguishable questions. EFAMA presents evidence of normal trading capacity and market maturation. ESMA tests the implications for asset eligibility and investor protection under adverse conditions. Each proposition can be substantively meaningful at the same time. The correct research response is to obtain conditional block-trade and fund cash-flow evidence capable of testing both claims.

Cash budgeting and liquidation arithmetic

A dated cash budget can be expressed as opening unencumbered cash plus certain receipts by deadline T plus executable sale proceeds settling by T plus committed funding available by T, less redemptions and other payments due by T. A positive portfolio net asset value is not itself cash. Proceeds expected after T cannot fund a payment at T unless a legally and operationally verified bridge exists. Principal subject to a possible catastrophe event cannot be assumed to repay at full face without analysis of the conditions.

Consider a wholly hypothetical $100 million fund facing $12 million of redemptions due in five business days. It has $4 million available cash and $2 million certain note maturities settling before the deadline. It must raise $6 million more. If an identified portfolio block can actually sell at 93.75 percent of current principal and complete settlement in time, selling $6.4 million notional produces $6 million before costs. This is a valid calculation only under its assumptions. The price must be firm for the entire notional and available before the redemption deadline.

Suppose the actual sale ladder instead provides 93.75 for the first $2 million and 91.00 for the following $4.4 million, at the same current-principal basis. Gross proceeds equal $1.875 million plus $4.004 million, or $5.879 million, leaving a $121,000 shortfall before fees and settlement adjustments. The apparent portfolio wealth may remain large even as the immediate cash constraint binds. Selling an additional note, retaining more cash, or accepting a legally available liquidity-management tool would change the outcome, but no generic remedy may be assumed available without checking the fund's rules.

The worked example also illustrates why simple quoted spreads are inadequate. If the manager uses a 93.75 bid without checking depth, the liquidity budget may incorrectly show no shortfall. Execution quantities and settlement times are part of the cash-flow forecast, not ancillary trading details.

Fair valuation, NAV, and model governance

A portfolio may be valued using executed trades, indicative broker marks, multiple dealer quotations, a pricing service, or a model incorporating expected catastrophe loss. Those sources differ in timeliness and verifiability. A recent small trade may give a strong price observation for a small lot but be unrepresentative of a large sale. A model can provide a coherent conditional valuation but remain untested against current firm bids. A composite price may look independent while relying on the same underlying model inputs as other published sheets.

An internal valuation review should preserve timestamps, bid-versus-mid convention, quotation source, trade size, contractual principal base, model version, event forecasts, and records of overrides. Independent challenge should question stale marks, changes in expected loss, highly concentrated sponsor or peril exposures, and a common data provider across nominally distinct quotes. A mechanical spread between an arbitrary model value and an arbitrary small-lot bid should not be called a precisely estimated liquidity premium.

Aon's market evidence on executable demand demonstrates the possibility of strong bids relative to reference marks in a favorable market; ESMA's discussion of stressed tradeability demonstrates the importance of asking whether those bids persist during severe market events. EFAMA's account of specialist dealer quotations provides institutional information about the development of normal-state price discovery. Together they support disciplined price-source selection instead of a blanket assumption about either permanent illiquidity or guaranteed depth.

Stress testing and governance

A robust stress exercise should vary the catastrophe's location and severity; estimated principal impairment; investor withdrawal intensity; secondary quote withdrawal and dealer capacity; settlement timing; new issuance competition; and correlations of fund investor behavior. Each scenario should calculate a dated cash shortfall and the weighted-average liquidation concession, then show the remaining portfolio after forced asset sales. High-quality stress scenarios use actual dealing rules, contractual note terms, and tested price-quantity ladders.

Model risk deserves equal prominence. A very severe hazard model can lower the note's conditional expected value without introducing unusual selling pressure. Adding an arbitrary “illiquidity haircut” on top of a model output that already reflects forced-sale prices double counts the same effect. A second model may produce a different event loss estimate; the valuation process should display this uncertainty separately from the execution-cost assumption.

The Swiss Re 2026 issuance review and Aon 2026 market growth figures show that new investor capital and growing market scale provide supportive liquidity conditions in certain periods. A credible stress exercise must also reverse those conditions and require several sellers to liquidate the same peril simultaneously. Normal-market growth is valuable context, not a substitute for downside cash-flow analysis.

Integrated Comparative Synthesis and Research Implications

Findings that survive methodological comparison

The source record establishes three conclusions. Outstanding catastrophe bonds do trade in a specialist secondary market, with observable dealer transactions, quoted valuations, and active institutional participation. Executing or anticipating a trade involves measurable costs; the best empirical studies report a material liquidity component of required returns. Execution capacity is conditional on instrument, quantity, event state, and investor flows, so annual turnover and total outstanding value do not establish universal liquidity.

The direct and proxy studies support a disciplined interpretation. Zhao and Yu's historical liquidity proxy estimates show liquidity compensation under an indirect estimation method. Herrmann and Hibbeln's realized secondary trade evidence measures transaction-related friction in a selected security population. Braun's issuance-pricing evidence and Froot's catastrophe-risk capital constraints provide complementary capital-market context. The different studies test connected questions; their coefficients should not be pooled mechanically into one number or described as a single causal trend.

Industry figures require equal care. Aon's June 2026 outstanding estimate reports $63.4 billion. EFAMA's later 2026 market estimate reports approximately $65.6 billion. HANetf's October 2026 quarterly report, using Plenum Investments data, gives a Q3 figure of roughly $65.5 billion and reports a secondary risk-spread compression from approximately 5.71 percent near June 2026 to 4.57 percent near September. These are market estimates at different observation dates and under potentially distinct inclusion rules; their numerical proximity cannot be used to infer a single audited global total. Changes in a risk-spread index are not, without additional information, changes in the bid–ask execution spread.

The Artemis ILS report series corroborates the usefulness of separating issuance activity and outstanding stock from trading activity. A high issuance quarter can create new investment alternatives and shift relative valuations even where the total investor capital pool is increasing. Analysts should specify whether a quoted figure concerns newly issued limits, current outstanding principal, cash trading volume, capital raised in funds, or a market-value index.

International and contractual variation

Market participation is global, yet individual securities may be concentrated in selected perils and countries. Aon's 2026 geographic issuance breakdown indicates a very high share of North American peril exposure in the recent market period. Such concentration matters because investors diversified by sponsor can still own highly correlated event exposure. Following a major North American catastrophe, several otherwise independent portfolios may simultaneously approach their exposure limits, reducing the effective buyer pool for the same risk.

The World Bank Jamaica transaction shows an alternative sovereign sponsor context with parametric disaster protection. It documents a real risk-financing contract and distinctive event trigger, but publication of the issuance terms alone does not prove the existence of continuous two-way secondary quotes. The same caution applies to indemnity, industry-index, and modeled-loss notes. Trigger design affects information arrival; specific trade evidence determines liquidity.

A simplified cross-asset turnover ranking should also be handled carefully. Corporate bonds and catastrophe bonds differ in typical maturity, loss drivers, investor bases, reporting coverage, and dealer activity. EFAMA's descriptive cross-bond-market turnover comparison can inform discussion but cannot eliminate selection and measurement differences. Acharya and Pedersen's liquidity-risk compensation logic supplies a conceptual basis for testing common risk, not license to transplant coefficients from equities to catastrophe bonds.

Evidence hierarchy and future research

A research-grade secondary-liquidity database would link bond-level original terms and current outstanding principal to timestamped executed trades, firm and indicative quote histories, offered quantities, unsuccessful trade inquiries where lawfully obtainable, expected loss model versions, catastrophe event developments, and relevant fund investor-flow data. The unit of analysis would be declared. An interval with no executed trade would remain an explicit no-trade observation, rather than be filled with a fabricated price.

The principal outcomes should include realized execution cost, bid–ask spread at stated size, effective execution probability, days or hours to completion, price impact, quote resilience, and fund cash shortfall by due date. Empirical designs could control for instrument fixed effects, peril and event-period interactions, remaining maturity, issue size, security structure, and macro-financial liquidity. Strong causal conclusions would require credible exogenous variation and a convincing treatment of endogenous trade selection. Even if a natural disaster is physically external to a dealer, security exposure and portfolio selling decisions can be endogenous.

Robustness exercises should test alternative trade-pair matching windows, quotation conventions, size caps, late and revised event estimates, missing eligible securities, reference-rate definitions, and risk-model sensitivity. An analyst should always reconcile source populations and observation windows before comparing numerical premiums. A full source-to-claim ledger makes it possible to distinguish the primary empirical record from policy arguments, market promotion, and hypothetical examples.

Original analytical synthesis

A scientifically meaningful liquidity assessment starts with a concrete execution proposition: can a named investor sell a defined outstanding principal amount of a specified security, in a particular event-information state, before a specified cash deadline, for a price no worse than an explicitly chosen concession? Every relevant statistic then has an identified role. Executed transactions support claims about realized prices. Firm quotes and quoted sizes support claims about conditional depth. Turnover supports claims about historical activity. Models support estimates of expected contractual loss. Dated fund cash-flow budgets support claims about redemption capacity.

These are complementary evidence classes rather than interchangeable proxies. A conclusion about full-fund cash sufficiency demands more information than a historical trade count; a conclusion about the market price of event risk demands more information than an instantaneous dealer spread. ESMA's supervisory concerns and EFAMA's description of ordinary trading can therefore coexist without logical contradiction. Their empirical resolution requires event-conditional transaction depth and dated funding obligations.

This synthesis is an original interpretation of the reviewed literature, contractual arrangements, and institutions. It applies established microstructure and financial analysis to the catastrophe-bond context. It makes no claim to a new authored doctrine, proprietary model, statistical estimation, or universal regulatory finding.

Conclusion

Catastrophe-bond secondary-market liquidity is the practical ability to transfer an outstanding catastrophe-linked note for cash under defined conditions. Its determinants include contractual payment and loss rules, information about events, specialist intermediary networks, institutional investor mandates, issue size, order imbalance, valuation methods, and the time available for execution. Historical tradeability and normal-period turnover provide real evidence of market function, while the ability to raise a large sum quickly during correlated catastrophe and redemption stress remains a distinct and more demanding question.

Zhao and Yu (2019) estimate a liquidity component equal to 9.42 percent of average spreads under their proxy-based historical design. Herrmann and Hibbeln (2023) estimate approximately 21 percent under a separate transaction-based design. Both establish substantive liquidity compensation within stated samples. Their coefficients cannot be used interchangeably or treated as a timeless market-wide haircut. Aon, Swiss Re, EFAMA, HANetf, and Artemis document the expansion and functioning of the market through 2026 without providing an unconditional executable bid for every note and scenario.

The regulatory comparison reinforces the same conclusion. ESMA's 2025 technical advice concerns eligibility and stressed liquidity; EFAMA's November 2026 study documents trading capacity and argues for institutional accessibility. An evidence-based assessment respects each document's status and tests the specific claims through observed quantities, prices, and fund liabilities. In practice, catastrophe-loss revaluation, liquidity premiums, quoted bid–ask costs, execution depth, and redemption cash budgets must be measured as related but distinct objects.

The monograph's original contribution is a source-grounded comparative synthesis, a careful distinction among measurement concepts, and transparent hypothetical execution and liquidity-budget calculations. The central conclusion is precise: secondary liquidity in catastrophe bonds is a time-specific, size-specific, contract-specific, and event-conditional capacity to execute. Its scientific measurement requires identified securities and market states, verified data provenance, and explicit recognition of what each source can and cannot demonstrate.

References