This paper analyzes a sequential approach to lifting interventions in the COVID-19 pandemic taking heterogeneity in the population into account. The population is heterogeneous in terms of the consequences of infection (need for hospitalization and critical care, and mortality) and in terms of labor force participation. Splitting the population in two groups by age, a less affected younger group that is more likely to work, and a more affected older group less likely to work, and lifting

more »

... ntions sequentially (for the younger group first and the older group later on) can substantially reduce mortality, demands on the health care system, and the economic cost of interventions. Abstract This paper analyzes a sequential approach to lifting interventions in the COVID-19 pandemic taking heterogeneity in the population into account. The population is heterogeneous in terms of the consequences of infection (need for hospitalization and critical care, and mortality) and in terms of labor force participation. Splitting the population in two groups by age, a less affected younger group that is more likely to work, and a more affected older group less likely to work, and lifting interventions sequentially (for the younger group first and the older group later on) can substantially reduce mortality, demands on the health care system, and the economic cost of interventions. In a more likely scenario, the epidemic spread will rebound as interventions are eased, especially given the global interconnectedness of the world economy. An evaluation of approaches to progressively lifting interventions and switching from suppression to mitigation strategies is hence urgently needed. We focus on the health and economic consequences of a sequential approach that lifts interventions for the less vulnerable fraction of the population first and for the more vulnerable fraction of the population second. We show that this approach reduces mortality while increasing economic activity (relative to delayed lifting of interventions) by allowing the fraction of the population, which participates in the labor market to a greater extent, to return to work earlier. That said, technological progress in testing and pharmaceutical interventions, including development of treatments and a vaccine, continue to be a priority, as is increasing the capacity of the health system in terms of testing, hospital beds, and critical care. In fact, a mitigation approach makes such progress more urgent not less. An added benefit of imposing non-pharmaceutical interventions on the more vulnerable population for longer is that, by the time interventions are lifted on this population, health system capacity has increased and treatments, and even a vaccine, may be available. This paper considers a deliberately simple discrete time version of the Kermack and McKendrick (1927) Susceptible-Infected-Recovered (SIR) model with a heterogeneous population to study the implications for mortality, the demands on the health system, and the economy. We consider a population comprised of two groups (but the approach can be extended to multiple groups in a straightforward matter). The two populations differ in the consequences of infection and in their participation in labor force. Specifically, one group is comprised of younger people who are less affected by infection and have a higher labor force participation rate. The other group is comprised of older people who are more affected by infection and have a lower labor force participation rate. We show that a sequential approach to lifting interventions, in which the less affected younger group is released earlier and the more affected older group is released later, can substantially reduce mortality, the demands on the health care system, and the drop in economic activity. In our baseline specification, mortality is reduced by close to 40% and peak load hospitalizations and critical care demand by about 75% and 80%, respectively, compared to lifting interventions on both groups at the same time. Mortality is reduced because by the time restrictions are lifted for the more vulnerable population a sizable part of the population has recovered reducing the infectiousness of the pandemic. A staggered lifting of interventions reduces peak load demand for health care directly, but importantly also indirectly, if the more vulnerable group is released only once a sizable fraction of the