Prioritizing Testing and Workplace Re-opening Under COVID-19

Click here for a table of Michigan industries detailing proximity, ability to work from home, layoff risk and employment.

The Issue:

States are contemplating and initiating the reopening of their economies at a time when the virus has not completely disappeared and months before the earliest vaccine is even possible. This requires the need to balance the economic losses of households and businesses against the danger of resurgence of the virus. Finding the safest way to open and operate businesses is a key part of this balance. However, protective gear and testing remain in limited supply. One alternative could be to use specific industry characteristics to guide industry openings in a way that lowers contagion risks and maximizes economic benefits while broader testing becomes available and the efficacy and reliability of immunity testing is ascertained.

Prioritizing for testing and reopening occupations in which workers can operate at greater distance could reduce contagion and maximize benefits.

The Facts:

• Though widespread use of masks and an effective vaccine are essential for complete reopening, the supply of masks remains limited and effective vaccines are months away. Many recent policy recommendations have outlined the steps necessary to reopening, emphasizing widespread use of testing to find and isolate any cases of active infection and the need to equip workplaces with personal protective equipment (PPE) — such as face masks — to reduce contagion from asymptomatic carriers (see here and here, for example). Currently, PPE supplies remain limited even for first responders and healthcare providers. Widespread diversion of these resources towards non-essential uses is impractical. Though vaccines are in development, their implementation has an uncertain timeline. The use of immunity testing, which identifies people who have recovered from the infection and are presumably no longer vulnerable, has also been proposed as a way to find workers able to safely return to work. Although the CDC issued new guidelines allowing essential workers who have been exposed to COVID19 to return to work, there remains uncertainty around the nature of immunity. As we develop understanding of the virus and the human response to infection, immunity testing could be combined with effective use of PPE and physical distancing to minimize the spread of the virus. By doing so, a more nuanced approach to social distancing can be achieved enabling an earlier implementation of mechanisms to restart production and jump start the economy. 
• Some occupations and industries are better able to operate with effective physical distancing and this translates into regional differences in the types of exposure to contagion through work. To capture differences in physical distancing among occupations, we use the metric of proximity index (PI) recently developed by St. Louis Federal Reserve based on the Bureau of Labor Statistics O*NET database. The PI number indicates the physical distance from other people involved in carrying out work by occupation: an index below 50 indicates a low-contact occupation such as scientists, designers and hunters; high-contact occupations have PI numbers above 75 and involve arms-length contact or near touching such as barbers, hairstylists and cosmetologists, many health care workers, food and beverage servers and also pilots, air traffic controllers and flight attendants (see here). The differences in contact intensity across occupations translate into regional and geographic differences across the country in exposure to COVID-19 impacts through work, due to geographic differences in the distribution of production. We constructed the PI for Detroit metropolitan statistical area (MSA) to illustrate differences among occupations and the numbers of workers involved in different levels of contact-intensity work. Due to the importance of the auto manufacturing industry in the region, workers in transportation and equipment manufacturing comprise an important share of those employed and they tend to have lower contact-intensity occupations: for instance, assemblers and fabricators (PI: 58) and mechanical engineers (PI: 58) have a lower proximity index because they work in office or factory floor settings. In contrast, food preparation and service workers (PI: 94) and health care workers (PI: 88) have among the highest proximity measures because they involve having direct physical contact with others.
• Low-proximity industries tend to interact predominantly with other low-proximity industries, reducing the economy wide footprint of viral spread. No industry operates in a vacuum; production requires inputs from other industries; consumption of the final product of the industry may involve high proximity. Evaluating the input and consumption patterns of industries at the national level, a recent study finds that 92.1% of inputs in low-proximity industries are purchased from other low-proximity industries.  Furthermore, 42.6% of production of low-proximity industries serves as input in other low-proximity industries, with only 16.2% reaching high-proximity industries (see here). These estimates suggest that though low-proximity industries are not isolated, their operation is largely interdependent on other low-proximity industries, making targeted testing and reopening in these industries a viable option.  
• Some occupations can work from home while others cannot. Given the essential nature of some occupations (e.g., healthcare workers, first responders), which have a high PI, resumption of activity cannot be based simply on prioritizing occupations with low-proximity index. (Although there is room for innovation in some areas that could increase distance even in occupations that are traditionally high contact. For example, telemedicine has transformed medical interactions under stay-at-home orders, allowing patients to receive care at a distance from traditionally high proximity occupations.) In some occupations, such as accounting (PI: 46), though adequate physical distance can be maintained at work, a large part of work can be carried out using technologies from home. This means that during current lockdown conditions, firms providing professional services, such as accounting, can operate at up to 74 percent of their capacity as their workforce continues to be productive from home. The same cannot be said of factory production.
• Industries with factory production cannot operate from home, resulting in higher likelihood of layoffs. To illustrate the interaction of proximity with operational capacity, we aggregated occupations by industry and performed analysis of the automotive industry in Michigan (see chart). Automotive is one of the largest industries in Michigan, employing over 188,000 people, with an industry-wide PI of 59.  The industry employs a wide variety of occupations, some of which can be performed from home such as designers and engineers (PI: 51), while others depend on factory production and thus cannot be performed from home (e.g. assembly and fabricator). Though 65.3% of workers in design and engineering can work from home, workers in production cannot do so. This means that during this lockdown, the automotive industry as a whole operates at up to 16.7% capacity. For production, the automotive industry in Michigan employs approximately 124,000 workers who are, as a result, at risk of furlough and layoffs.
• Widespread furlough and layoffs can have a devastating impact on a local economy.  Based on the average monthly income of occupations in the automotive industry, we estimate that layoffs would result in $495 million loss in personal income for each month of lockdown. In addition to the stress on household finance, this loss of income translates into lost tax revenue and business activity for communities in Michigan. Low private savings in socioeconomically vulnerable communities, such as Detroit, will magnify the impact of lost personal income. In the longer term, reduced tax revenue will undermine delivery of local social services, such as food pantries, already straining under increased use.  
• Immunity testing of workers in targeted low-proximity low-capacity industries could eventually be an efficient and practical use of a limited resource. Reopening an industry, even one with a low PI and high potential for gain in operational capacity, may precipitate a resurgence in the spread of COVID-19. Therefore, if immunity from the virus from prior exposure is demonstrated, we could balance resumption of economic activity and infection risk using immunity testing for industries based on the former but seeking to mitigate the latter once reliable antibody testing is developed and becomes widely available. Developing combined approaches with nasal swabbing for viral infection followed by blood immunity testing could identify individuals who are not acutely infectious and are either developing immunity or fully convalescent. The latter group would be well-positioned to return to work, scaling implementation based on the regional impact of a given industry, within the context of potential for interpersonal proximity should immunity not be complete. 
• Potential wintertime recurrence of COVID19 may necessitate industry prioritization. Though economic activity is reacting immediately and sharply to social distancing, a recent study projects recurrence of wintertime outbreaks of severe acute respiratory syndrome transmissions, necessitating intermittent social distancing and economic disruption through 2022. As many states are preparing to reopen their economies, future recurrence will necessitate similar, though milder, mitigation measures. Prioritization of industries now will reduce the economic fallout of future social distancing measures.