という点について分析したNBER論文が上がっている。論文のタイトルは「Inequality of Fear and Self-Quarantine: Is There a Trade-off between GDP and Public Health?」で、著者はSangmin Aum(明知大学校)、Sang Yoon (Tim) Lee(ロンドン大学クイーン・メアリー校 )、Yongseok Shin(セントルイスワシントン大学)。

Our model provides a framework for quantitative analysis and can be used to evaluate and predict the aggregate and distributive effects of real-world policies in various economic settings. The quantitative nature of our analysis sets it apart from most other works in the literature, which tend to be either empirical or theoretical. In this paper, we choose the model parameters to replicate the progression of COVID-19 in South Korea and the United Kingdom (henceforth SK and UK, respectively). These two make an interesting and informative contrast. SK responded early with aggressive testing and tracking, and largely contained the epidemic. The UK on the other hand belatedly imposed a blanket lock-down, and its containment efforts have not been as successful.
Based on our quantitative analysis of the two countries, we obtain three key results.
First, contrary to the common view, there may not be a clear trade-off between GDP and public health after all. It is true that, since a lock-down prevents people from working normally, it can slow down the rise in infection at the expense of lower economic output. It is also true that a premature lifting of the lock-down increases GDP initially at the expense of rising infections. However, if the lock-down is lifted too soon, infections can rise to a level at which most people voluntarily work from home out of fear of infection, and this would happen in a matter of months. The government may try to impose another round of lock-downs, but all the countermeasures lose their potency once infections reach a certain threshold. For the UK, an extended lock-down will lead to 150,000 fewer infections and 5-percent higher GDP by November than the current policy. In other words, a stricter and longer lock-down can deliver both higher GDP and better public health outcomes.
Second, if the UK had adopted the SK policies, its GDP loss would have been smaller by two-thirds both in the short and the long run, and the cumulative infections through November would have been smaller by 70 percent. This is not merely because SK implemented policies sooner: The model shows that an earlier implementation of the lock-down in the UK has minor effects on GDP and infection. Rather, it is because aggressive testing and tracking can more effectively isolate the infected and hence reduce their chance of infecting other people, without forcing everyone, including the large majority that is not infected, to work from home where they are less productive.
Third, the epidemic or the policies implemented to counter it do not affect people equally. Low-skill jobs tend to be more contact-intensive (e.g., restaurants and retail), which means that (i) low-skill individuals face higher infection risks and hence suffer more from the fear of infection, and (ii) it is hard to do their work from home and hence their earnings loss when working from home is larger. For these reasons, low-skill workers and self-employed are disproportionately affected by the epidemic and the government's countermeasures that make them work from home (be it through testing, tracking and/or lock-down), and some low-skill workers in particular switch jobs in response. One exception is the potential policy of issuing "virus visas" to those who have antibodies: This policy will disproportionately benefit the low-skill workers and self-employed, by relieving them of the fear of infection and also by allowing them to get back to work. A visa policy can raise UK's GDP by 10 percent compared to its baseline lock-down policy in our model, entirely driven by a 20-percent higher output in low-wage sectors. This result suggests that antibody tests should prioritize the low-skilled, not only to help those most in need, but also to have a maximal positive effect on aggregate GDP.