Infectious diseases, externalities and uncertainty
Epidemics and/or pandemics involve market failure, which requires public interventions. The most commonly argued causes of market failure surrounding infectious diseases are externalities – one would be more likely to contract the virus when other people have contracted the virus (i.e. negative externality), and conversely, actions such as avoiding contact with other people would benefit other people (i.e. positive externality).
The current COVID-19 crisis is however strongly suggesting another cause of market failure, which urgently requires public interventions – an uncertainty that takes time to resolve and has no objective assessment, where the latter aspect is often referred to as `Knightian uncertainty’ (Knight, 1921). We usually make an explicit distinction between before and after the resolution of uncertainty (i.e. the distinction between ex ante and ex post). In contrast, the current crisis is a dramatic demonstration that the length of time until the resolution of uncertainty matters crucially. In particular, the key uncertainty is that no one knows how long the pandemic will last, and this feature differentiates pandemics from other catastrophes such as earthquakes and floods (although their effects may extend far into the future). This means that we need to seriously consider ex interim in addition to ex ante and ex post.
Coping with the time dimension
Part of the uncertainty we are currently facing (i.e. the ex-interim uncertainty), is about the duration of the pandemic. Many authorities around the world are currently imposing extremely strict, nonpharmaceutical interventions (or community mitigation strategies) that drastically restrict economic activities, and no one knows how long the current suspension of economic activities will last. There is clearly a trade-off between human casualties and the economy, but there may well be significant health effects or even casualties caused by the economic suspension unless a public financial intervention is introduced to counter the financial hardships people are facing.
Although the current nonpharmaceutical interventions may be relaxed depending on how the spread of the virus slows (and indeed some countries have started cautious, partial relaxing of measures), it will not be easy politically and/or socially since no reliable and objective assessments of different strategies will be available until a cure and/or a vaccine is developed. The difficulty of relaxing nonpharmaceutical interventions is evidenced by existing studies – for instance, Hatchett et al. (2007) studied nonpharmaceutical interventions in different cities in the US during the Spanish flu pandemic, and found that they were effective in smoothing the peak when they were introduced at an earlier stage, but that relaxing the interventions caused renewed viral spread.
There is a familiar risk in which the time dimension is central – longevity risk (i.e. no one knows for sure when one will die). In countering longevity risk, pensions play a crucial role, and a similar measure is needed to counter the financial difficulties many people are currently facing arising from the current economic suspension. Specifically, rather than giving a one-off payment, payments should be given on a rolling basis until the pandemic is over, and financing such schemes may well require very large public debt issues if the current pandemic lasts for longer than a year. This scheme is effectively a pension scheme functioning in reverse, in which payments will be given first, and the contributions will be made later through taxation (cross-subsidies through progressive taxation will also be necessary).
How can we better prepare for future pandemics?
The fact that experts have been warning about the possibility of a global pandemic shows that the current crisis was not completely unanticipated. A global pandemic belongs to either a class of Knightian uncertainty that is called `ambiguity’ or a class that is called `unforeseen contingencies’. Neither of these classes of uncertainty allows for assessing future outbreaks probabilistically. Thus, a standard cost-benefit analysis is unsuitable here. While existing studies such as James and Sargent (2006) found that the aggregate economic impacts of past pandemics were usually smaller than initially predicted, aggregate measures such as GDP or aggregate (monetary) losses are inappropriate, since such measures ignore distributional effects and effects on human casualties and health or on future generations through prolonged school closures. Instead, the maximin principle or the Rawlsian criterion should be applied here, i.e. to seek the best worst case.
To prepare for future pandemics, we must learn from the current crisis so that we will be able to achieve the best worst case. A possible policy is to allow for excess medical capacity or slackness in terms of both personnel and equipment, which a standard cost-benefit analysis would reject, since the latter focuses on the average rather than the worst case. Also, the very nature of the uncertainty surrounding a new infectious disease makes early detection inherently difficult. This calls for a mechanism that incentivises active international co-operation, allowing detected cases to be reported immediately and information to be shared globally from the very outset.
April 17, 2020