The pandemic abolishes the diversification on which reinsurance lives, by striking all policyholders at once. After the Covid bill and the failure of pandemic bonds, the risk can be borne only by a public-private partnership, which mutualizes in time what cannot be diversified in space.
Insurance rests on a silent assumption, that losses strike policyholders independently of one another. A fire, an accident, a water-damage event hit one member of the pool without touching the others, and it is this independence that allows the law of large numbers to stabilize the average cost and to make the risk insurable. The pandemic violates this assumption in the most radical way possible, for it does not strike one policyholder or one region, it strikes by definition almost all the members of the pool at the same time1. Where a natural catastrophe ravages one territory and spares the others, the pathogen spreads to the whole globe, annihilating the separation of risks on which insurance is built. The pandemic is therefore not one large risk among others, it is a risk of a different nature.
This peculiarity strikes reinsurance at its heart, for reinsurance lives precisely on diversification. Its trade consists in pooling uncorrelated risks, a Japanese earthquake with a Florida hurricane and an Australian drought, on the reasonable bet that these events will not all occur on the same day. Geographical and sectoral diversification is its raw material, what allows it to bear colossal commitments with relatively modest capital. Yet the pandemic destroys this resource, since it simultaneously correlates all lines, life insurance through excess mortality, business insurance through interruption, event insurance through cancellations, and all geographies at once. There is then no longer any elsewhere over which to spread the risk, no uncorrelated portfolio to cushion the shock. The pandemic does not merely exceed the market's capacity, it abolishes the very mechanism by which it works.
The Covid-19 crisis turned this theoretical difficulty into a very real bill. According to the broker Howden, global insured losses linked to the pandemic reached around 44 billion dollars, making it the third costliest claim in the history of insurance, behind only Hurricane Katrina and the attacks of 11 September 20012. The initial projections, which spoke of more than 100 billion dollars, fortunately did not materialize, largely because many policies already excluded the risk or contested it. The shock was nonetheless felt across the whole chain. The reinsurer Swiss Re thus set aside 2.5 billion dollars in the first half of 2020, before adding further reserves, and recorded a net loss where it would have generated a profit of 1.6 billion without the pandemic, its property combined ratio jumping above 110%3. For a player of this size, used to absorbing major hurricanes without faltering, such a reversal of result under the effect of a single peril has the value of a warning.
The singularity of this claim lies less in its amount than in its provenance. The bill was built up by the addition of lines that nothing usually connected, excess mortality in life insurance, cancellations in event insurance, business interruption in commercial insurance, credit and directors' liability. The case of Wimbledon became emblematic, for the tournament's organizers had purchased pandemic coverage within their cancellation insurance for seventeen years, for an annual premium of two million dollars, and the cancellation of the 2020 edition earned them a payout estimated at around 141 million dollars4. This rare foresight stood as an exception, and its cost to the insurer illustrated what the coverage of such a risk represents when it is actually granted. Most other players, for their part, had not anticipated the simultaneous accumulation of all these lines.
One part of the claim remained more discreet in public debate, although it touches the very heart of reinsurance, that of excess mortality. Life reinsurers bear, through mortality treaties, the risk of an excess of deaths relative to the actuarial tables, and a pandemic is precisely the event that realizes this risk on a large scale. Covid-19 caused marked excess mortality in several countries, in particular among the elderly and the working-age brackets in the United States, weighing on the benefits of life and protection contracts. This dimension recalls that the pandemic strikes not only the economy through business interruption, but also the insurers' balance sheet through mortality itself, and that it thus correlates two risks that everything separated, the economic risk and the biometric risk, usually lodged in distinct lines and assumed to be independent.
The main battlefield was that of business interruption. Classically, this coverage indemnifies the operating loss following physical damage, a fire or a flood that prevents the business from functioning. Covid-19 raised an unprecedented question, that of whether an interruption without physical damage, caused by an administrative closure or a lockdown measure, could also trigger the coverage. Everything then turned on the interpretation of clauses, including those relating to infectious diseases and to the prevention of access to premises. In the United Kingdom, the regulator brought a test case before the courts in order to settle quickly and for the whole market, and the Supreme Court delivered in January 2021 a ruling largely favorable to policyholders in the case opposing the FCA to Arch Insurance and several others5.
The contrast between jurisdictions was striking. In the United States, where more than twelve hundred proceedings were brought, the courts mostly dismissed policyholders on the ground that no direct physical damage was characterized, and the sector argued that the pandemic was by nature uninsurable6. Nearly 80% of American property policies moreover already carried, since 2006, an exclusion targeting viruses and bacteria. The lesson of this battle joins the one already encountered in relation to hybrid war, that the ambiguity of a clause is paid for dearly. When the drafting does not explicitly anticipate a risk, the doubt benefits the policyholder, and the insurer finds itself exposed to an accumulation it had neither wanted nor priced. Covid-19 thus revealed a silent exposure to pandemic risk, concealed in wordings designed for other perils.
Beyond the very existence of the coverage, a second front opened on the question of aggregation. When a clause covered the interruption following an infectious disease occurring within a certain radius around the premises, was the pandemic to be treated as a single event, capped only once, or as a multitude of local events whose caps added up. The financial stake of this characterization was considerable, for it could multiply the insurer's exposure by the number of recognized infectious clusters. British case law adopted a reading favorable to policyholders on causation, holding that the pandemic and the governmental measures formed a sufficient concurrent cause, which further increased the bill for the market. These subtleties of seemingly technical appearance determined billions in indemnities, and illustrate how far the pandemic tests the finest architecture of contracts.
The market's reaction was not slow in coming, and it took the form of a massive retreat. To prevent such an ambiguity from recurring, the Lloyd's Market Association published model exclusion clauses for communicable diseases, broken down by line, LMA5393 for direct property policies, LMA5394 for the corresponding reinsurance treaties and LMA5395 for liability7. These clauses, rapidly incorporated into contracts, expelled pandemic risk from standard coverage, acknowledging that business interruption of pandemic origin was no longer insurable by the private market. The sector thus clarified its position, but at the price of a widening of the protection gap, since businesses now find themselves without coverage against a peril whose reality the crisis had just demonstrated.
This retreat is not a passing timidity, it follows from the fact that the pandemic fails simultaneously several of the conditions of insurability. The risk must be independent from one policyholder to another, estimable from a stable history, geographically diversifiable and of a magnitude compatible with the market's capital. The pandemic contradicts each of these criteria at once, which distinguishes it from ordinary catastrophes, which generally violate only one8.
Faced with the incapacity of the traditional market, many pinned a hope on transferring the risk to financial markets, in the image of catastrophe bonds. The idea was not new, since Swiss Re had issued as early as 2003 a first extreme-mortality bond of 400 million dollars, triggerable if a mortality index in five countries exceeded 130% of its reference value9. The most ambitious attempt was the Pandemic Emergency Financing Facility, launched by the World Bank in 2017 after the Ebola outbreak, endowed with around 425 million dollars of capacity under its insurance window, and intended to mobilize funds rapidly for the poorest countries10. These bonds transferred to investors the risk of losing their capital if a pandemic crossed predefined thresholds.
The Covid-19 experience cruelly revealed the weakness of this mechanism. The World Bank's bonds were triggered only in April 2020, that is around forty days after the WHO had characterized the situation as a pandemic, and paid out only 195.84 million dollars, a derisory sum compared with the 160 billion the Bank finally mobilized by other means10. The trigger criteria, judged too late, too complex and deliberately restrictive, had been calibrated to reduce the probability of payment, to the point that investors had received 96 million dollars in interest before the write-down. The facility was not renewed. This failure illustrates the pitfall specific to parametric instruments applied to the pandemic, namely the definition of the trigger, which joins the attribution problem already encountered for hybrid war. Determining the precise moment when a pandemic begins, and the threshold from which it gives rise to payment, is an exercise in which every trade-off moves the indemnity away from the real need.
From this twofold observation, the incapacity of the traditional market and the limits of transfer to capital, follows a conclusion that most players now share. Pandemic risk, like war risk and like extreme climate risk, can be borne only within a public-private partnership, where the state intervenes as insurer of last resort11. Several schemes have been proposed on the model of the existing arrangements for terrorism, whether a British Pandemic Re inspired by Pool Re, the American bill on pandemic risk insurance, or government-backed shared funds in Europe. All rest on the same intuition, only the community has a base broad enough to mutualize a risk that no private portfolio can diversify.
Designing such a scheme nonetheless raises difficulties that it would be wrong to underestimate. One must first define the triggering event, and the declaration of a pandemic by the WHO, while offering a convenient marker, does not necessarily coincide with the moment when the economic losses become massive, which reproduces the parametric pitfall at the state level. One must then allocate the burden between insurers and the public authority, entrusting to the market the tranches it can absorb and to the state the truly systemic share, in the image of the layered architecture of Pool Re. One must finally contain moral hazard, by preventing an overly generous public guarantee from discouraging businesses from any prevention. None of these difficulties is prohibitive, but their resolution conditions the viability of a market rebuilt on mixed foundations, and explains why no major country has yet instituted a lasting scheme several years after the crisis.
The pandemic nonetheless presents a singularity that distinguishes it even from war, and that reinforces the necessity of public recourse. The magnitude of the operating losses was not determined by the virus alone, but by the lockdown and closure decisions taken by the public authorities. The state is thus, to a large extent, the author of the claim it would be called upon to cover, which makes illusory the idea of a purely exogenous risk that the market could price independently of political action. This entanglement of the risk and the public decision argues for the state to assume directly a share of the burden, since it partly controls its occurrence and its magnitude. Where the insurer can neither foresee nor influence lockdowns, the public authority decides them, and it would be incoherent to transfer to it the cause of the claim without also transferring its cost.
At the end of this analysis, the answer to the initial question, after Covid what market, is sober. There no longer exists, and there can scarcely exist, an autonomous private market for the pandemic, for the risk abolishes the diversification on which insurance lives. What the state can offer and the market cannot is a pooling not in space, impossible for a global peril, but in time, by spreading the burden of a pandemic over several generations of taxpayers through recourse to debt. The state's capacity to borrow against the future is precisely what allows it to bear a risk that, being everywhere at the same instant, can be spread nowhere in the present.
This substitution of temporal pooling for spatial pooling is not neutral, for it shifts the burden from today's policyholders to tomorrow's taxpayers, and turns a question of insurance into a question of justice between generations. It also supposes that the state retains a sufficient borrowing capacity at the moment the pandemic strikes, a condition that is not guaranteed for the most fragile economies, which are also the most exposed. The public-private partnership is therefore not a miracle solution, but the least bad of arrangements in the face of a risk that defies the logic of insurance, and its institution will depend as much on political choice as on actuarial technique.
The reinsurance of pandemics is therefore not so much a market to rebuild as a frontier to recognize, the one beyond which the private transfer of risk gives way to collective solidarity, a frontier this series has already encountered for climate and for war, and which takes shape each time a risk becomes truly systemic12.
The war exclusion rested on an identifiable war. Hybrid war erases declaration, attribution, boundaries and the kinetic character, making the exclusion almost inapplicable. From NotPetya to Lloyd's retreat, how to insure a risk designed to remain without an author.
The Insurance-Linked Securities market represents a quiet revolution in global finance: for the first time, catastrophe risk migrates off insurer balance sheets into the portfolios of pension funds and institutional investors.
Underground economy, ransom markets and insurer pricing: how ransomware became a structured industry whose growth is partly financed by cyber insurance payouts.
A power grid goes dark, a hospital freezes under a ransom. Crime, disaster, or act of war. On the answer hangs everything, because if it is crime your policy pays, and if it is war it does not. And those who launched the attack have arranged that no one can tell.
A structure falls in a town you have never heard of, and a retiree's savings in Tokyo wobble that same month. A hidden wire runs between the two, and that wire is reinsurance.
One analysis with every new publication, on insurance, reinsurance, cyber and AI. No account, no noise.
No spam, one-click unsubscribe, no data ever sold.