How do catastrophe bonds work?

The sponsor identifies a risk to transfer

A catastrophe bond starts with a sponsor, typically an insurer, reinsurer, or government entity, that wants to transfer a specific catastrophe risk off its balance sheet. The sponsor defines the peril (hurricane, earthquake, flood), the geographic region, the coverage period, and the amount of protection needed.

This is not a general transfer of all risk. Cat bonds are structured around a narrowly defined set of conditions laid out in the bond's term sheet. The sponsor works with arrangers and modelling firms to quantify the risk and structure a deal that will be attractive to investors.

A special purpose vehicle is created

The sponsor does not issue the bond directly. Instead, a special purpose vehicle is set up, usually domiciled in an offshore jurisdiction like Bermuda or the Cayman Islands. The SPV sits between the sponsor and investors. Its sole purpose is to issue the cat bond notes and manage the collateral.

This structure insulates investors from the sponsor's credit risk. Even if the sponsor were to become insolvent, the collateral held by the SPV remains protected in a trust account. It also provides tax and regulatory efficiency for both parties.

Investors buy the notes and collateral is locked

The SPV issues notes to investors, who are typically institutional players such as dedicated ILS funds, hedge funds, pension funds, and asset managers. Investors pay the face value of the notes, and those proceeds are deposited into a collateral trust account.

The collateral is usually invested in low-risk instruments like money market funds or Treasury bills to preserve capital. This is a fully collateralised structure, which means the money to pay potential claims is already set aside from day one. There is no counterparty credit risk in the way traditional reinsurance carries it.

The sponsor pays a premium, investors receive coupons

Throughout the life of the bond, the sponsor pays a regular premium to the SPV. This premium, combined with the return earned on the collateral, funds the coupon payments that investors receive. Coupons are typically quoted as a spread above a reference rate.

The spread compensates investors for the catastrophe risk they are taking on. Higher-risk bonds, those with higher expected losses or lower attachment points, pay wider spreads. The relationship between expected loss and spread is one of the most closely watched metrics in the cat bond market.

The trigger determines the outcome

The trigger is the mechanism that decides whether investors lose principal. If a qualifying catastrophe event occurs and meets the trigger conditions, some or all of the collateral is released to the sponsor. If no trigger event occurs during the risk period, investors receive their full principal back at maturity.

There are four main trigger types. Indemnity triggers are based on the sponsor's actual losses. Industry loss triggers are based on market-wide losses reported by an index provider. Modelled loss triggers use catastrophe model output applied to a defined exposure set. Parametric triggers use a physical measurement like wind speed, earthquake magnitude, or rainfall depth at specified locations.

Each trigger type carries different trade-offs. Indemnity triggers give the sponsor the closest match to their actual losses but take longer to settle and require disclosure. Parametric triggers settle quickly but introduce basis risk, the possibility that the trigger pays out when the sponsor has no loss, or does not pay out when the sponsor does have a loss.

Maturity or loss

Most cat bonds run for three to four years. At the end of the risk period, if no trigger event has occurred, the collateral is released from the trust and returned to investors along with the final coupon payment. The SPV is wound down and the transaction is complete.

If a trigger event does occur, the loss calculation determines how much collateral is released to the sponsor. Some bonds have binary triggers where the full principal is lost if the threshold is breached. Others have a sliding scale between an attachment point, where losses begin, and an exhaustion point, where the full principal is consumed. Losses between these two points are shared proportionally.