Alternative Investments

Catastrophe Bonds and Insurance-Linked Securities

Insurance-linked securities transfer insurance risk, usually natural catastrophe risk, from insurers and reinsurers to capital market investors. In a catastrophe bond, investors fund a collateral account and receive a spread above the collateral yield; if a defined catastrophe exceeds the attachment point, part or all of the principal is used to pay the sponsor's losses.

What are insurance-linked securities?

Insurance-linked securities are financial instruments whose returns depend on insurance outcomes rather than on corporate earnings or interest rates alone. The most common are catastrophe bonds, which cover losses from events such as hurricanes, earthquakes and severe storms. The category also includes collateralized reinsurance, industry loss warranties, sidecars that share in a reinsurer's portfolio, and longevity or mortality linked instruments. The sponsor, typically an insurer, reinsurer, government entity or large corporation, pays a premium to transfer a defined risk, and investors earn that premium in exchange for agreeing to absorb losses if the event occurs.

For sponsors, the market provides capacity beyond traditional reinsurance, multiyear coverage and collateral that eliminates counterparty credit risk. For investors, it offers a return driven largely by natural events, which have little direct connection to economic cycles. That low correlation is the main reason institutional portfolios allocate to the asset class, though it is not a guarantee of diversification in every scenario. During some periods, such as broad market stress, prices of insurance-linked securities can still fall alongside other risk assets as investors seek liquidity, even when no catastrophe has occurred.

How is a catastrophe bond structured?

A catastrophe bond is issued by a special purpose insurer or reinsurer that enters into a reinsurance or retrocession agreement with the sponsor. Investors buy notes issued by the vehicle, and the proceeds are placed in a collateral trust invested in high quality, short dated assets such as government money market funds. Investors receive interest equal to the return on the collateral plus a risk spread funded by the sponsor's premium. If no qualifying event occurs during the term, typically several years, investors receive their principal back at maturity.

If a qualifying event occurs and losses exceed the attachment point, collateral is released to the sponsor to pay its covered losses, and investors lose that portion of principal. Losses between the attachment and exhaustion points are usually allocated proportionally, so a bond might lose part of its principal rather than all of it. Because the collateral is segregated and invested conservatively, investors face very little credit risk to the sponsor; their main risk is the insured event itself. Coupons therefore float with short term rates through the collateral return.

• Sponsor pays premium to the special purpose vehicle
• Investors fund collateral held in trust
• Investors receive collateral yield plus risk spread
• Principal is released to the sponsor after a qualifying event above the attachment point

What trigger types are used?

The trigger defines what loss counts. An indemnity trigger pays based on the sponsor's actual losses, which aligns coverage closely with the sponsor's exposure but requires investors to rely on the sponsor's claims handling and portfolio disclosure, and settlement can take longer. An industry index trigger pays based on estimated insured industry losses from an event, as reported by an independent agency, which is more transparent but leaves the sponsor with basis risk if its own losses differ from the industry. A parametric trigger pays based on physical measurements such as wind speed or earthquake magnitude at defined locations, which settles quickly and objectively but can have significant basis risk.

Modeled loss triggers run event parameters through an agreed catastrophe model applied to a notional portfolio. Each trigger type balances transparency for investors against precision of coverage for the sponsor. Investors usually demand more spread for indemnity triggers on less transparent portfolios, and sponsors accept parametric or index triggers when faster payment or lower cost outweighs basis risk. Understanding the trigger is the first step in any analysis, because the same modeled event can produce very different payouts under different trigger designs, and disputes over loss determination can delay settlement.

How are attachment, exhaustion and expected loss measured?

The attachment point is the loss level at which the bond begins to take losses, and the exhaustion point is where principal is fully consumed. Catastrophe models estimate the annual probability that losses reach each point. The probability of attachment is the chance the bond suffers any loss in a year, and the probability of exhaustion is the chance it loses everything. The expected loss is the average annual principal loss across all modeled scenarios, which lies between these two probabilities.

Investors compare the spread with the modeled expected loss. The ratio, often called the multiple, indicates how much compensation the bond pays per unit of modeled loss. A higher multiple suggests more cushion for model error and for the cost of tying up capital. Model risk is substantial, because catastrophe models rely on limited historical data, evolving building stock, exposure concentrations and changing climate patterns, so prudent investors consider results from more than one model and their own views of loss trends.

What risks do investors face in insurance-linked securities?

The core risk is that a covered event occurs and principal is lost. Beyond that, investors face model risk, since modeled probabilities can understate true risk; correlation risk, because a single large event can hit many bonds at once, especially if a portfolio is concentrated in one peril and region; and extension risk, because principal can be held past maturity while losses from a recent event are being determined. Liquidity in the secondary market is limited and can disappear when a major storm is approaching. Collateral risk is small but not zero, and depends on the quality and structure of the trust investments.

Investors also face pricing cycles. After large loss years, spreads tend to widen as capital leaves and sponsors pay more for protection. After quiet years, new capital enters and spreads compress, which lowers compensation for the same risk. Disciplined investors track the multiple over expected loss rather than headline yields to judge whether pricing is adequate. Comparing the current multiple with its history for similar perils and layers provides a useful signal about whether new issuance is attractively priced. Investors should also note changes in terms during soft markets, such as broader covered perils or weaker reporting, which can quietly increase risk without any change in headline spread.

How should an allocator underwrite an ILS portfolio?

An allocator should evaluate an insurance-linked portfolio by peril and region exposure, expected loss, the distribution of possible annual losses including tail scenarios, trigger mix, sponsor quality and the manager's modeling approach. The key output is not average return but the loss in a severe year, such as a one in one hundred or one in two hundred year event, relative to the portfolio's annual income. Diversification across perils such as hurricanes, earthquakes and European windstorms reduces volatility, but some regions dominate the market, so true diversification is harder to achieve than the number of bonds suggests. The allocator should size the position so that a severe loss year is tolerable within the total portfolio.

Principal loss and pricing for an industry index catastrophe bond

1. A $100,000,000 catastrophe bond attaches at $1.5 billion of industry insured losses and exhausts at $2.0 billion, with losses allocated linearly across the $500 million layer.
2. An event causes $1.8 billion of industry losses. The layer loss is ($1.8 billion - $1.5 billion) / $0.5 billion = 60%, so investors lose $100,000,000 x 60% = $60,000,000 of principal.
3. Modeled probability of attachment is 2.0% and probability of exhaustion is 1.0%, so expected loss is approximately (2.0% + 1.0%) / 2 = 1.5% per year.
4. The bond pays a 6.0% risk spread. Multiple = 6.0% / 1.5% = 4.0x expected loss.
5. If the collateral yields 4.0%, the total coupon is 4.0% + 6.0% = 10.0%. In a year with no event, investors earn 10.0%; the expected return after modeled losses is about 10.0% - 1.5% = 8.5%.

How underwriters apply this

• Read the trigger definition carefully and assess basis risk and settlement timing for each type.
• Compare risk spread with modeled expected loss and track the multiple across market cycles.
• Use more than one catastrophe model and adjust for exposure growth and climate trends.
• Measure portfolio losses in severe tail years, not just expected annual loss.
• Limit concentration by peril and region, recognizing that some exposures dominate the market.
• Review collateral trust investments and extension provisions after events.

Frequently asked questions

Keep reading

Educational content written by practitioners. Not investment, accounting, legal or tax advice.