How Economists Read Models of Coronavirus Epidemics

Unprecedented events create new demands for understanding. If they do occur, we reach out to any model that can help us understand them and how to respond. This is the case with the novel coronavirus, which relies heavily on epidemiological models to predict how the disease will spread and the effectiveness of various measures that can be taken to stop it. Dealing with viruses has also challenged existing economic models that were not built in light of the types of shocks that are occurring today. This includes simultaneous disruption of both supply and demand.

Models are essential in very uncertain situations. Also, excessive dependence is dangerous. Without them we are blind, but with them there is a risk of blindness. They can focus ourselves on the role of the factors that are taken into account, ignoring those that are not, and lead us to an explanatory cul-de-sac. The actions that policymakers consider may be limited to the actions expected in the model, and some actions may be ignored because they were considered unmodeled or not modelable.

Economists are familiar with the issues involved, and so should others. The model becomes an empty castle, which is exquisite and beautiful, but lacks the quality of reality. “Calibration” according to parameters that appear to match the real world can lead to incorrect accuracy. It may be because the true value is uncertain. In the case of coronavirus, its mortality rate is still controversial. Or maybe it’s because you don’t completely know the relationships between the various variables. For example, it’s unclear how sick the disease will spread among people who use the park instead of indoor spaces for shared recreation, requiring closure of the park and banning outdoor exercise. To act in the world, you need to have a model that represents it appropriately, use that model to reach certain conclusions, and then apply them to the world. However, the world-to-model mapping and vice versa can result in many deviations.

As an economist, I am not trained in epidemiology, but I am trained in understanding and evaluating models. The most influential model used by government everywhere to guide the response to the coronavirus is by a team led by Neil Ferguson, an epidemiologist at Imperial College London who advises the British government. The created model. Produced empire model Big quote Number of potential deaths in the US and UK early in the crisis Stressed Closing schools and businesses as a response.

You will never know if the estimates of possible deaths in the absence of lockdowns are correct, as they were actually performed. But what can be said is that the model was limited in various ways. The health effects of policy options assessed beyond coronavirus mortality are not considered. These effects have already spread around the world, from individuals who have not been screened or treated for cancer to antimalarial treatment and vaccination of children who are not provided. The impact of non-health outcomes, such as economic consequences including unemployment and poverty, was not considered.

Imperial model Premises For example, it was very detailed and empirically informed in certain areas, given how the population is spatially distributed. However, they seem to have been drawn from the thin air of other people, such as envisioning what percentage of the infection will come from a “community contagion”, rather than a family, school, or workplace. Another example was that it was assumed that children were more than twice as likely to transmit the disease as adults, which might help to justify school closure, but The data suggests If anything, it is unlikely to do so. The model did not take into account the potential for great success elsewhere. Taiwan Schools remained open during the early SARS epidemic, but interactions between children were limited by eliminating breaks and other means. It also goes beyond the general idea of ​​reducing social contact for all seniors and considers measures to create guards to prevent diseases from reaching them in nursing homes and elsewhere. I didn’t. Similarly, the model assumed only partial compliance with social distance and workplace closure requirements, but the level of compliance was Much higher. As the model envisions, humans reasonably respond to threat knowledge, not just as randomly moving molecules. If policy makers were more aware of the detailed assumptions driving model outcomes, they could The decision they made It was more cautious about implementing other options or just more positive about the premise of its choice.

One thing that seems to be lacking awareness of how the conclusions depend on assumptions, or at least not being recognized by public authorities, and acting in accordance with the advice of scientific authorities, is one thing. Emphasized as if to mean. Details of other models used to guide the decision, such as the model created by the University of Washington by the University of Washington Seems to have used Recently, no transparency has been done to generate predictions of disease trajectories.

Models need good judgment, but good judgment is also necessary. In order for it to be applied, it is essential to recognize that the model is constructed in different ways and reflects a range of plausible assumptions. Lack of complete understanding of the novelty of the phenomenon (in this case, a properly named new coronavirus) and the complex human factors involved in carrying it, inevitable judgment, scientific knowledge For both respect, is very uncertain and therefore inadequate.

Politicians, and society in general, can’t leave scientists with the heavier decisions to be made than scientists can give politicians their briefs. Care must be taken with science, but wise recognition of the extent of uncertainty also applies. Similarly, attention must be paid to the diverse purposes that need to be balanced in a society where properties such as safety and freedom are often very different. As attention shifts to the content of possible exit strategies from lockdowns that create the complex problem of balancing different values, the proper role science plays in the public square is not a replacement for it. , Even more important.

Sanjay G. Reddy is an economist at New Social for Research in New York. You can follow him on Twitter. Yuta.