This article was first published on March 4 but Covid-19 case data is being updated every day.
March 5 update: Corrected to remove additional requirement for an IDA country to have suffered at least 20 deaths before being able to apply for aid. Also updates to include details of growth rate and confirmation ratio calculations, and updates Euromoney assessment of earliest trigger date to April 6.
March 6 update: Latest clarifications from World Bank: confirmation that start of event was determined to be December 31, 2019, and further trigger requirement that at least one IBRD/IDA country be affected.
March 11 update: Confirmation of growth rate and confirmation ratio assessment period running until April 6. Two days allowed for calculation, so EARLIEST REPORTING DATE OF WHETHER ALL BOND TRIGGER CRITERIA MET IS APRIL 9.
A World Health Organization (WHO) situation report on Sunday March 22 catalogued 292,142 cases of coronavirus Covid-19 across 187 countries and territories, causing 12,784 deaths.
In 2017 the World Bank issued a pandemic bond designed to help fund the response to any widespread outbreak of a number of diseases, including coronavirus. The $320 million bond was part of a bigger $425 million risk transfer that included a concurrent $105 million swap with six reinsurance counterparties.
Should it already be paying out?
As of March 16, the answer was no, principally because the required time period has not yet elapsed since the start of the outbreak, which is 84 days, or 12 weeks. The World Bank confirmed to Euromoney on March 5 that the official start of the outbreak has been set at December 31, 2019, meaning that the first day following the 12-week period will be tomorrow, March 24.
Contrary to rumours circulating online, there was no requirement for a body like the World Bank or the World Health Organisation to formally declare that the outbreak is a pandemic for the bonds to pay out. When the criteria are met, the bonds will be triggered, whether or not a pandemic has been declared. However, the WHO did in any case declare a pandemic as of March 11.
Some of the bond trigger criteria have been met, but two - the growth rate of cases and the ratio of confirmed cases to total cases (including suspected) - can only be calculated after the initial 84 days, and are determined during an additional period of up to two weeks.
The growth rate calculation is particularly problematic: in the case of coronavirus, it is calculated on the basis of individual IBRD or IDA countries only, meaning that it is highly dependent on those countries' ability to test and report cases.
It is also difficult to assess externally because it uses for its calculation the total case amount, which includes probable cases that are not routinely reported publicly by the WHO in its situation reports.
The timeframe means that although March 24 marks the end of the 12-week period since the start of the outbreak, the growth rate and confirmation ratio data that relate to that day would not be known until April 6. Two days is then set aside for calculation of those rates.
On that basis, Euromoney understands that the earliest date on which it would be known if all bond trigger conditions have been met is April 9.
If the scheme does pay out, it will do so only to a very targeted group of countries – the poorest in the world, which are also those with the weakest healthcare systems and likely to be the slowest to begin to report accurate assessments of an outbreak.
Structuring the pandemic bond, as Euromoney reported in depth at the time, was immensely complicated, not least because it needed to strike a balance between being useful and being investible.
The bond will mature on July 15, although it can be extended by one year. The World Bank is preparing a second iteration of the scheme, dubbed Pandemic Emergency Facility (PEF) 2.0, which the Bank said last year was to be marketed in May 2020.
Some of the triggers depend on cases of disease reported by developing countries – those that fall within the World Bank's lending categories of IBRD (International Bank for Reconstruction and Development) or IDA (International Development Association).
Among the parameters governing whether or not the bond is triggered, and how much it pays out if it is, are: the number of countries affected; the number of cases in each of those countries; the number of deaths; the percentage of confirmed cases to total cases, including suspected; and the growth rate of cases.
All of these will be assessed based on specific reporting periods. The conditions to trigger the bond need to be in place at least 12 weeks after the designated start of the event for payouts to happen. After that, they must be in place on a rolling 12-week basis.
In the event that the bond is triggered, it will release funding only to IDA countries. A country seeking assistance is required to request the funds, although requests can also be made by certain designated responding agencies, chiefly multilateral development banks and United Nations agencies.
One hurdle, then, is a country's ability to reliably record and report cases and deaths. At the moment, 27 IDA countries have reported confirmed cases of coronavirus, according to the WHO's data on March 15, with 121 cases and one death.
According to a data from the Africa Centres for Disease Control and Prevention on March 18, every country in Africa apart from Mali, South Sudan and Botswana had reported either confirmed or suspected cases.
Since the pandemic bond is intended to help developing countries deal with serious outbreaks of disease, there has been criticism that it was not triggered by outbreaks of Ebola virus since its creation, including an outbreak in the Democratic Republic of the Congo last year. The complexity of the trigger mechanisms and the resulting difficulty in modelling the payout scenarios have also come under fire.
However, while the bonds were not triggered by last year's Ebola outbreak, the companion cash window scheme that forms the other part of the PEF was triggered and paid out $50 million to the DRC. The World Bank committed a further $350 million to that outbreak, via the IDA.
But all that has now been dwarfed by the World Bank's announcement on March 3 of an immediate $12 billion emergency aid effort to assist developing countries in tackling the coronavirus Covid-19 outbreak.
So far in the current coronavirus outbreak, the pandemic bond has not been triggered. What needs to happen for that to change?
Below are the key triggers for the ‘insurance window’ (the funding provided by the pandemic bond) to pay out.
1) Time period
A period of 12 weeks has to pass from the designated start of the outbreak, as marked by an Eligibility Event Notice. The World Bank has told Euromoney that an Eligible Event Notice was issued by the IBRD on January 31, and also confirmed on March 5 that the start date of the event was determined to be December 31, 2019.
That means the earliest date on which the criteria for triggering the bonds could begin to be assessed would be March 24. However, this is not the end of the story when it comes to the trigger date. Two of the most complex criteria - the growth rate in new cases and the confirmation ratio of confirmed cases to total cases - are based on data for the period beginning at the start of the week containing the day being assessed, through to the end of the following week.
That takes the date to April 6, Euromoney has now learned that there is also a two-day period allocated for final calculation of the rates, meaning that the earliest date when the status of all triggers could be known would be April 9.
2) Cases and geographic spread
At that point, the outbreak must be affecting more than one country anywhere in the world. If up to seven countries are affected, it is designated as a regional outbreak; if the number is eight or more, then it is designated as global. For a country to count towards this total, it must have at least 20 confirmed deaths.
In addition, the World Bank confirmed to Euromoney on March 5 that the outbreak must be occurring in at least one IBRD or IDA country, a condition that was first met by China.
The rolling total number of cases specifically in IBRD/IDA countries – in other words, the number within a rolling 12-week period – must be at least 250, and it must also still be growing at a rate determined by a complex formula.
The rolling confirmed number of cases must be at least 20% of the rolling total of cases, which includes those merely suspected to be coronavirus (see point 5 below for more details on the confirmation ratio).
As of the March 22 data from the WHO, there are now 17 countries globally with at least 20 deaths, meaning that the outbreak would now be designated as global.
In addition to the country-level requirement for 20 deaths for the assessment of a regional or global pandemic, there must be a total of at least 250 confirmed deaths in IBRD/IDA countries in order for the scheme to be triggered. With 3,267 reported as of March 22 in China alone, the 250-death requirement has already been met for the current 12-week period.
The theoretical payout to IDA countries of the $195.8 million that is available in the case of coronavirus is staggered based on the number of total deaths in IBRD/IDA countries. At 250, the payout is 29% ($56 million) for regional outbreaks and 34% ($66 million) for global; at 750 this rises to 57% ($113 million) for regional and 67% ($131 million) for global.
The maximum requirement of 2,500 deaths has already been met, meaning that the full regional/global payout of $195.8 million is available, assuming all other criteria are met. There are formulas for determining the exact level of funding that each country can receive, based on factors including population size.
4) Pandemic bondholder losses
The World Bank’s pandemic bond was structured with two classes. The $95 million class-B tranche pays out first, and can be reduced to zero by a coronavirus payout. There is a greater hurdle for payout from the $225 million class-A tranche, and the loss here is capped at 16.67%, or $37.5 million.
At 250 IBRD/IDA deaths within the time frame, and assuming the rolling case and growth rate requirements have been met, the class-B tranche suffers principal reduction of 37.5% in the case of a regional outbreak and 43.75% in the event of a global outbreak. At 750 deaths, this rises to a 75% loss for regional and 87.5% for global. At 2,500 deaths the entire tranche is lost.
The rolling deaths hurdle for the class-A tranche is 2,500, which has already been met, meaning that as things stand, the class-B tranche would be completely wiped out and the class-A tranche would suffer its maximum principal reduction of 16.67%.
In late February, the class-B bonds were marked at about 45, while the class-A were at between 90 and par.
The difference between the maximum $195.8 million assistance available to countries in the case of coronavirus and the maximum bondholder losses of $132.5 million is accounted for by the concurrent reinsurance swap.
5) Growth rate and confirmation ratio
These are the hardest to assess externally, and impossible ahead of the time at which most of the criteria would become applicable on March 24 - the reason being that both requirements are to be assessed based on the data from the two-week period in which the day occurs when all other requirements are met.
In other words, if all other criteria were satisfied on any day after the 12-week period, the growth rate of cases would be calculated based on data for the period beginning at the start of the week in which that day occurred through to the end of the following week, so a maximum of 14 days.
Based on the end of the 12-week period being March 24, that would put the end of the earliest growth rate and confirmation ratio reporting period at April 6. There are then two days allocated for final calculations.
The growth rate calculation itself is complex. It is based on the growth in total cases in each of the IBRD/IDA countries affected, and that includes probable cases, which the WHO does not routinely report in its published Situation Reports. Additionally, it is subject to various statistical smoothing methodologies.
The result of the growth rate calculation must be greater than zero, which will be known on April 9 at the earliest.
The confirmation ratio, meanwhile, relates to the proportion of confirmed cases to total cases, including suspected. It is additionally difficult to ascertain externally as it relies on detailed insight into suspected cases reported by national authorities.
The confirmation ratio must be more than 20%, which also be known on April 9 at the earliest.
Cash window depleted
The World Bank's pandemic bond was the largest and most innovative part of a broader response programme set up by the multinational organization, called the Pandemic Emergency Facility (PEF).
It was implemented in response to what was widely recognised within the World Bank and the World Health Organization to have been a slow and ineffective reaction to a bad outbreak of Ebola in 2014.
The PEF has two parts: a cash window and an insurance window. The cash window, which is intended to be sized at $50 million to $100 million, is funded by donor countries and has more lenient triggers than the insurance window.
The insurance window is $425 million and was funded by capital markets issuance – the $320 million bond and the $105 million swap. Bond coupons are paid for by the PEF (passed through to donor governments) and by the World Bank.
The cash window should, in theory, pay out before the insurance window. For coronavirus, it can be triggered by 30 cases in an IDA country where human-to-human transmission can be confirmed, and 100 cases if not. The cases must have occurred within any four-week period within the six weeks prior to a request for assistance being made.
However, while the cash window was envisaged to be replenished each year, the World Bank has told Euromoney that as of March 2 it had not yet been replenished in 2020, meaning that a mere $2.7 million is currently available.
Since being put in place, the PEF’s cash window has been used three times – all to help fight Ebola. In May 2018 it contributed $11.4 million to the Ebola response in the Democratic Republic of Congo, which also received two further tranches of $20 million and $30 million for the outbreak in 2019.