The current coronavirus outbreak started in late December 2019 from China, but the World Health Organization (WHO) declared a so-called Public Health Emergency of International Concerns (PHEIC) one month later, only after its director-general met with Chinese President Xi Jinping. Many have found suspicious this apparent excessive deference toward the Chinese government, which raised serious doubts about the agency’s decision to delay the alarm. Suspicions such as these are not new, however. WHO was criticized for downplaying the risks at the onset of the 2014 Ebola outbreak. The agency was also accused of exaggerating the risk by declaring a PHEIC for the 2009 H1N1 epidemic, supposedly to boost the profits of the pharmaceutical industry.
Yet before accusing WHO of misbehaving - no definitive evidence of conflicts of interest have surfaced thus far – other explanations for these misrepresentations of the risk need to be considered. The most immediate one is that WHO simply misread the information it possessed at the time of these events. Assessing the seriousness of an outbreak is indeed difficult because early information is typically limited and ambiguous, especially for a new virus such as COVID-19. A second and possibly more intriguing explanation is that misrepresenting the risk may actually improve the effectiveness of WHO’s warning message over time. My co-authors and I have explored this question in a forthcoming paper in Management Science. In this work, we seek to determine why an agency such as WHO might have an incentive to distort the early signals it has about an outbreak. To that end, we develop and study a framework based on so-called dynamic Bayesian persuasion games.
Declaring a PHEIC is one of the most difficult decisions that WHO has to make. Besides the inherent difficulty to assess the seriousness of an outbreak, WHO has no mean to enforce the actions it might requests. Only its member states have the capability to do so. The agency can only rely on its credibility, perhaps its most important asset, to elicit a response. And even if member states trust the quality of WHO’s assessments, these countries may not have an incentive to act. Indeed, countries look after their own population and economy, while WHO’s mission is to protect the world. Crucially, member states and their populations update the agency’s credibility over time by comparing the agency’s warnings with the actual incidence of successive outbreaks. This means that failure to raise the alarm, or on the contrary ``crying wolf’’, may hamper the agency’s ability to elicit mitigation actions in the future.
Briefly, the results of our analysis indicate that the agency’s need to manage its credibility sometimes leads the agency to send warning messages that misrepresent its knowledge about an upcoming disaster. When its credibility is too low, the agency downplays the risk and actually downplays more as its credibility improves. By contrast, when its credibility is very high, the agency sometimes exaggerates the threat. In this case, the less credible the agency is, the more it exaggerates. Only when the agency’s credibility is somewhat moderate does it consistently send warning messages that fully disclose its private information about a potential disaster.
It is too early to say if our result explains the apparent delay in raising the PHEIC for the corona outbreak. Although we have corroborative evidences that the incentives we uncover are at play, only robust econometric studies will be able to confirm or falsify our findings. One thing is sure, though: different possible rationales for strategic information disclosure exist, and not all of them are bad for society. One should think them through before accusing the WHO of misbehavior when it raises the alarm.
Francis de Véricourt (ESMT Berlin)
S. Alizamir, F. de Véricourt, and S. Wang. 2019. Warning Against Recurring Risks: An Information Design Approach, forthcoming in Management Science.
