Demanding productivity standards are closely tracked. Fifteen-minute breaks are barely long enough to reach the warehouse bathroom. An influx of robots work to eventually replace you. Vending machines distribute free pain medications to reduce the line outside the nurse’s office. Investigative reporting suggests serious injury rates are nearly twice the national average.
It hardly came as a surprise when the Seattle Times estimated Amazon’s U.S. employee turnover was roughly 100% during the first six months of the pandemic in 2020 — or two to three times the national average for its industry. California’s National Employment Law Project finds this rate of employee churn is rather typical for the company.
Amazon offers an extreme example, but work-related stress exacts a toll on society. It diminishes worker productivity, results in higher rates of illness, and leads to psychological harms that may infect entire households. For socially responsible investors, measuring the potential cost of work-related stress imposed on employees is critical to comprehensively evaluating a company. This can be accomplished by melding academic estimates of the specific costs of work-related stress with indicators of stress in a particular organization, such as the level of employee turnover.
Juliet Hassard and her colleagues systematically reviewed the literature on costs of work-related stress. The estimated annual cost varies widely across studies, from less than $100 per worker to more than $700. These figures are likely conservative. Most are limited to measuring the direct costs of medical treatment for stress-related illness and estimating productivity losses related to sick time, presenteeism (i.e., “at work” but functioning below a typical level due to stress), and early retirement. Intangible costs like a lower quality of life due to the psychological effects of work-related stress are rarely considered.
The same study notes that intangible costs often account for more than half of the total cost of a particular illness when included. The Health and Safety Executive of the United Kingdom estimates that intangible costs account for 53% of work-related illness, of which work-related stress is the second most common cause. This figure is derived from stated preference surveys of respondents’ willingness-to-pay to avoid illness or injury. Such an approach may result in biased estimates, but including some measure of intangible costs is better than ignoring them altogether. A more appropriate range for the annual cost of work-related stress for each worker, including intangible costs and updating estimates for the passage of time, might be roughly $250 to $1700.
Using this range as a starting point, the next task is to identify company metrics that might be indicative of stress levels within an organization. The results of employee satisfaction surveys may be helpful, but analysts will likely find it difficult to compare data across companies. A more objective and more promising measure is employee turnover. Employee turnover is a component of the Global Reporting Initiative (GRI) standards (GRI 401-1), so it may become an increasingly common disclosure in sustainability reports.
The relationship between work-related stress and employee turnover has been most clearly established by targeted industry studies, but national surveys demonstrate a link as well. The American Psychological Association’s (APA) annual workplace surveys indicate consistently lower levels of employee turnover among companies it deems “psychologically healthy workplaces.” Employees of these companies are also much more likely to report the availability of stress management resources at work. 66% of survey respondents at APA’s 2019 Psychologically Healthy Workplace Award winners believed their workplace had adequate stress management resources, compared with 41% of respondents among all companies surveyed. Annual employee turnover at the award-winning workplaces, at 17%, was less than half of the U.S. average for 2018 (44%).
Building upon the association between employee turnover and workplace stress, we can map a scale of employee turnover relative to an industry or national average onto the range of the estimated per employee cost of workplace stress referenced above. The average level of employee turnover in the U.S. was 45% in 2019; this could be mapped to the midpoint of our $250 to $1,700 range, or $975 per worker. For U.S. operations, the estimated cost of workplace stress would be the product of $975 and the ratio of the company’s employee turnover relative to the 45% national average. In cases of extremely high or low employee turnover, the endpoints of the cost of stress per worker range could be used instead.
Criticisms of this method might include the mix of voluntary and involuntary departures in employee turnover statistics, as well as the reality that stress is not the only factor inducing workers to quit. Company employee statistics that separate out voluntary departures would be preferable. But since the percentage of voluntary departures is consistently north of 60% economy wide, at least in the U.S., it is not clear that disaggregated company labor data would change the picture all that much.
The multitude of reasons to leave employment beyond stress is a potentially more serious shortcoming of using turnover to gauge the cost of workplace stress. Some jobs are stigmatized as “low status,” and workers may look to leave as soon a better opportunity comes along. Does this approach unfairly punish companies for offering low-wage, low-status work?
A worker will rationally leave a job when the perceived costs exceed the perceived benefits of staying put. If an employee is in a low-stress job that nevertheless is harmful to their self esteem, there is still an intangible cost being absorbed that may encourage them to quit. The likely result of this method is the capture of additional social costs beyond the effects of stress. I would argue this further recommends it, even if what we are measuring is a basket of social harms including stress.
Precisely measuring the cost of workplace stress isn’t the point. The goal is to better allocate financial capital by internalizing the social costs of employment. In many cases, these costs will be significant relative to the company’s financial value. FedEx is the largest U.S. employer that discloses its employee turnover, which ranged from 30% to 51% for the three fiscal years included in the company’s 2020 sustainability report. Using a three-year average, I estimate a cost of workplace stress at approximately $890 per worker. With a headcount of more than 440,000, the estimated total cost amounts to nearly $400 million — equivalent to nearly 10% of Fedex’s operating profit.
Unsurprisingly, many companies with very high employee turnover don’t disclose the specifics publicly. If reported rates of more than 100% turnover at Amazon warehouses and Walmart stores are accurate, it implies that social costs related to employment may be as much as one quarter of the operating profit of those businesses (excluding Amazon’s AWS). This single social factor, if measured, would likely alter the portfolios of many socially responsible investors. At a minimum, there is a clear need for investors to demand employment disclosures consistent with GRI.