People buy insurance because paying a regular premium to have money available in a disaster makes budgeting easier. Large organizations are no different. They often buy insurance for lots of different catastrophes.
A few years ago, the World Bank decided to buy pandemic insurance. Nobody is currently selling pandemic insurance, so the World Bank needed to create a process to make it happen. They did this by selling catastrophe bonds, or “cat bonds.” If an investor buys cat bonds, then that investor is essentially acting as an insurance company. The holders of the cat bonds get a regular payment—kind of like an insurance premium—which gives them a return on their investment that is higher than a normal bond. But if the catastrophe happens, they lose some or all of their investment, as if they were an insurance company paying out a claim.
This system was called the Insurance Window of the World Bank’s Pandemic Emergency Financing Facility. The “insurance premiums” are paid by Germany and Japan. If the “insurance claim” is collected, the World Bank gets the money, and they decide how to use it to help respond to the pandemic. The current cat bonds last until July 2020, so the pandemic insurance will end then unless more bonds are issued.
The bond, like any insurance contract, specifies very precisely in advance the conditions under which the investors (insurers) would lose money. In order to eliminate legal uncertainty and allow the risk to be accurately priced, the triggers for the claim—that is, the definition of the pandemic—need to be very specific and numeric. The definition cannot be anything like “whenever the WHO declares an event a public health emergency of international concern (PHEIC),” because that condition is based at least in part on human judgment. If investors knew that they would lose their money whenever an organization makes a judgment call, they would not want to buy the bond.
Given the recent developments in the Congo Ebola epidemic, and the financial needs in responding to that crisis, many people are interested in understanding what exactly makes the bond trigger. What follows is my synthesis of the prospectus of the cat bond that the World Bank sold to investors. (Please note that I am not an expert in financial contracts or reading the prospectuses of cat bonds. This is not financial advice, or any claim that the bonds will or will not trigger soon. I may have missed something important about the trigger from an investor’s perspective; this is meant as my best interpretation of that prospectus regarding when the money will be available for pandemic response.)
The World Bank sold 2 kinds of cat bonds, 1 for influenza and 1 for other pandemics. The one relevant to the Ebola epidemic is the second one, the Class B note. For these notes, a pandemic is defined as an event that:
is caused by coronavirus, filovirus (Ebola), Lassa fever, Rift Valley fever, or Crimean Congo hemorrhagic fever;
kills at least 250 people;
lasts at least 12 weeks;
has at least 250 new cases in the past 12 weeks;
has an increasing average number of new cases over the past 12 weeks; and
kills at least 20 people in a second country.
If all of those conditions are met, bond holders lose some or all of their money, depending on which disease happened and how bad it was. A coronavirus or Ebola pandemic that kills 2,500 or more people means they lose everything, and the World Bank gets $95 million. If no pandemic happens, they get 11.5% a year above the risk-free rate, so the donor countries are paying about $11 million a year for this insurance.
If the current Congo Ebola outbreak had been spread over more than 1 country, the bond would have already paid out, with investors losing 60% of their money and the World Bank getting $57 million to respond to the pandemic, because the epidemic would have killed more than 750 people but fewer than 2,500 (so far). But because it was mostly confined to a single country, it did not meet all the conditions.
It is possible that the cat bonds will not pay out even if the epidemic spreads to another country and kills more than 20 people there. If the average number of new cases is stable or decreasing, then the bond will not pay out. All of the triggers must be met at once on the same date.
As of early October 2019, when this was published, the weekly number of new cases had been mostly decreasing for almost two months. This means that, even if the WHO confirmed 20 deaths in a neighboring country such as Tanzania, the money would not be available. Even if the disease develops into an endemic regional disaster that kills a regular and steady number of people for years, while the public health response must spend hundreds of millions of dollars on containment to prevent a larger disaster, the insurance money still may not become available, based on the stated terms.
Of the 4 other events that have been declared a PHEIC, the bonds would have paid out for 2 of them. They would not have paid out for the 2014 polio declaration or the 2016 Zika declaration, because those viruses are not covered in the contract. The 2014 West Africa Ebola epidemic would have triggered a full payout of the Class B notes, and the 2009 swine flu would have triggered a payout of the Class A notes.
The Class A notes will pay out $225 million for a flu epidemic that either causes at least 5,000 confirmed cases in less than 6 weeks or causes at least 5,000 confirmed cases with the number of new cases growing at over 27% a week, but only if it is a novel influenza A virus or one that has not been a seasonal flu virus in the past 35 years. There are no requirements for deaths or geographic spread; it just pays out everything, assuming that the flu in question has been the subject of a WHO report describing it. Also, if there is a coronavirus that has killed more than 2,500 people (and wiped out the Class B notes), then one-sixth of the value of the Class A notes will pay out.
Given that this is a relatively new financial instrument and that many in the pandemic preparedness community are interested in learning more about it, it may be useful for the World Bank to publish a plain-language description of the parametric triggers of both the Class A and B cat bond notes in their insurance window, which provides more information than is available in this brief summary. It may also be helpful to publish such descriptions before future cat bonds are issued in order to make sure that conditions that would trigger the release of funding are broadly understood.