The future of pandemic financing: Trigger design and 2020 hindsight
by Conor Meenan, Lead Risk Finance Specialist, Centre for Disaster Protection
Medics in full safety gear at the entrance to a new isolation unit at the Connaught Hospital in Freetown, Sierra Leone. Photo: Simon Davis / Foreign, Commonwealth & Development Office (FCDO).
By 17 April, all seven of the World Bank’s Pandemic Emergency Financing Facility (PEF) payout conditions had been met, triggering the release of $195.84 million from its insurance window in response to the covid-19 pandemic.
The total payout from the PEF is now greater than the sum of public money invested in the facility—and the ambition to better plan for and fund pandemic risk has proven prescient. However, in the three short but turbulent years since the launch of the PEF in 2017, aspects of the facility have continued to receive criticism.
From a technical perspective the PEF was meticulously designed, used state-of-the art modelling, and considered a complex risk in a comprehensive way. But two recurrent challenges have emerged—namely concerning its speed and complexity.
Each of these issues can be traced at least in part to the ‘growth rate’ payout condition, one of seven criteria that collectively define the parametric trigger for the insurance window.
Speed
The growth rate requires an acceleration of newly reported cases over a 12-week period. During the early phases of a disease outbreak, reported case data is volatile—the more time elapses, the more stable the picture of the event, so 12-weeks may have been chosen to represent a good balance of speed and stability. The modelled dynamics of coronavirus disease outbreaks and risk analyses will also have influenced selection of this timing window.
In retrospect, the 12-week window looks slow in relation to the covid-19 outbreak. And while the World Bank reports that at the time of triggering there were 4,653 cases reported in IDA countries (0.62% of global cases), these were collectively reported from 59 countries, so on other measures, the signals of a severe multi-country crisis were clear much earlier.
The issue of timing is further compounded by an additional two-week calculation period, the payment schedule of the insurance contract, and the logistics of disbursing funds once they have been triggered. If future pandemic risk financing instruments are intended to support the early stages of an outbreak, then these timing challenges need to be very carefully considered.
Complexity
In the scheme of parametric trigger mechanics, the growth rate calculation is complex. Complexity improves the accuracy of some parametric triggers, but at the same time it can also make the trigger process more complicated, less understandable, and seemingly less transparent.
For pandemics, a certain level of complexity is needed to meet the challenges of modelling and monitoring the underlying risk. The WHO report data used as the basis for the growth rate calculation is uncertain and volatile; data cleaning and smoothing processes, while complicated, are to some extent necessary.
The issue with complex triggers is that they introduce more assumptions, and place a greater reliance on the accuracy of risk models and observation data. The more moving parts in a parametric trigger, the more potential points of failure.
So, while the growth rate condition didn’t stop the PEF triggering for covid-19, had the number of new cases reported by the WHO followed a slightly different path, then it might easily not have triggered. This may seem counter-intuitive, but it is to raise the point that just because an instrument triggered, it doesn’t mean the trigger design was perfect.
What is the PEF’s ‘insurance window’?
The PEF has a ‘cash window’ and an ‘insurance window’.
The cash window is a donor-funded pot of money that can be spent quickly and flexibly as an outbreak is emerging. Funds from the cash window are allocated according to pre-agreed governance processes, but ultimately decisions are made at the discretion of a steering body. Funds from the cash window were approved in May 2018 for the Ebola response in the Democratic Republic of Congo (DRC). Further funds were requested and approved in 2019.
The insurance window is a larger pot of money that consists of private sector funds that are intended to be released for ‘more infrequent, severe events.’ In return for putting US$320 million on pandemic standby for three years, private market investors require two things: a return on investment (‘insurance premium’); and a watertight, objective, and legally binding description of the circumstances under which money is taken out of the pot. During the three-year term, if the pre-agreed trigger thresholds are all simultaneously exceeded, payouts are made to the World Bank to give to IDA countries and key responders. If they are not met, then the investors take back their original money alongside the premiums and interest they have received in the interim. This relationship between the World Bank and the private markets is similar to the relationship between any buyer and seller of insurance.
So why include the growth rate criteria?
It is hard to know exactly why the growth rate condition was included in the trigger design as these decisions are made behind closed doors. Though we might speculate about the combination of factors ...
As with the other conditions, the growth rate exists to control the circumstances under which the insurance pays out. The PEF was designed to ‘fill the financing gap that occurs after the initial outbreak and before large-scale humanitarian relief assistance can be mobilized.’ The growth rate also provides a view of whether the disease outbreak is worsening or improving. This target timing window may have something to do with the growth rate.
Cost is always a factor in the design of disaster risk financing (DRF) instruments, and is likely to have played into the decisions surrounding the inclusion of the growth rate criteria. Its addition limits the scenarios where all payout trigger conditions are met, reducing the probability of a payout , and therefore also reducing the annual premium costs.
Was the growth rate a requirement of the investors? Did it make the PEF more attractive for the private markets, and in theory improve the investability and cost-efficiency of coverage? Probably not. The PEF was the first of its kind: there was no precedent for the growth rate so it wouldn’t have been missed had it not been included. If anything, it adds another layer of complexity to an already complicated trigger, arguably making it harder for investors to wrap their heads around whether they will lose money, and providing an excuse to potentially charge more for investing, not less.
Early lessons
In retrospect, the issues of timing and complexity that stem from the growth rate would seem to outweigh much benefit—though parametric trigger design is always going to be much easier with hindsight.
There are early lessons here, not only for the next generation of pandemic financing, but also for other initiatives that push the use of parametric triggers into newer and ever more complex territory.
The PEF marks an ambitious attempt to lay the foundations for the future of pandemic risk management. The complexity of the PEF trigger is telling of this ambition: it pushed the boundaries of parametric trigger design in DRF—but perhaps went a little too far, too quickly.
There is value in starting simply and growing to the limits of what is feasible. It is much riskier to start at the fringes of possibility and iterate backwards.
The public opinion of PEF is a lesson in itself. The nuances of the costs and benefits of parametric triggers, and the use of insurance in development more generally, are easily overlooked when the DRF doesn’t work as people think it should. Expectation setting is a core part of the trigger design process; a more active outwards discussion of the intended role for DRF, its value and risk of failure is also needed.
PEF is a reminder that all disaster risk financing should be designed as if the world is watching.
The Centre’s Glossary of Terms can be found here.