The Great Resignation and the Paycheck Protection Program

Quits from Small Firms

It is informative to examine the breakdown of the aggregate quits rates by firms of different sizes using data from the Bureau of Labor Statistics (BLS). Quits are reported separately for six different sizes of firms based on numbers of employees 1–9, 10–49, 50–249, 250–999, 1,000–4,999, and 5000 or more workers. Firms that employ fewer than 500 workers are conventionally considered “small” businesses. Since the BLS classification for this series does not readily provide data below this threshold of 500, we will consider an alternative breakdown for understanding small versus large firms. We aggregate the three smallest sizes of firms and refer to this group as the “small firms,” that is, all firms that employ fewer than 250 workers. In contrast, the “large firm” group will be all firms that employ 1,000 or more workers.

Figure 2 presents the level of quits since January 2020 for small and large firms using the groupings just described, normalizing the level for January 2020 to 100 for ease of comparison. At the onset of the COVID-19 pandemic until May 2020, quits collapsed at both small and large firms, declining by almost 50 percent for small firms and 32 percent for large firms. As the labor market started to recover, quits recovered to their prepandemic levels and increased beyond that for small businesses, while large firms continued to experience quits that were in line with prepandemic levels. Hence, this figure highlights an important feature of the aggregate quits data during the Great Resignation period: Despite the presence of a large and sudden aggregate shock, small firms behaved differently than large firms and were primarily responsible for the excess quits seen during the subsequent recovery from the pandemic recession. Note that this divergence between firms of different sizes seems to be specific to the current recovery and was not a feature of the Great Recession (see Figure 3).

Figure 2 suggests that small firms as a group not only experienced drastic declines in their quit rates during the early phase of the pandemic but also accounted for the excess quits observed during the recovery. Since quits serve as a mechanism for the labor market to reallocate workers across different firms, unprecedented declines like those experienced during the beginning of the pandemic could have created a pent-up demand for reallocation during the recovery phase. However, based on the data depicted in Figures 2 and 3, this appears to be the case mainly for small businesses in the current episode. This pattern suggests that the key to understanding the drivers of the Great Resignation requires carefully examining reallocation patterns at small businesses specifically.

Quits and the Paycheck Protection Program

One potential reason for this pattern we see in the data might be related to an important policy response designed to address the effects of the pandemic on small businesses early on, namely, the Paycheck Protection Program (PPP). This program was enacted into law by the US Congress and was one of the largest elements of the Coronavirus Aid, Relief, and Economic Security (CARES) Act in 2020.³ Even though it was initially designed as a government-backed loan program for small businesses, the PPP effectively became a grant program for recipient firms as long as the grants received were used as the program intended and appropriate records were submitted.⁴ Therefore, the PPP could be interpreted as a government subsidy for small businesses to keep their payroll levels constant in the face of the uncertainty presented by the pandemic and the virus transmission mitigation efforts instituted by state, federal, and local authorities. This program might explain why small businesses experienced relatively low levels of separations including quits near the onset of the pandemic. It is also conceivable that as the labor market improved, more workers found themselves working in jobs that normally would have been eliminated in the absence of the PPP. In other words, the PPP might have prevented some quits at small firms early on, either through job elimination or the lack of opportunity for workers to find different jobs that better aligned with their needs; however, during the recovery the previously low levels of quits and the increased opportunity for workers to obtain jobs that better align with their needs would have resulted in excess quits at these small firms beyond what historical patterns of reallocation would have predicted.

We evaluate this hypothesis using JOLTS data for quits at the industry level. Unfortunately, JOLTS does not provide publicly available data on quits with a joint breakdown by industry and firm size. However, industry-level utilization of PPP loans provides useful information on variation over time. We can use these differences both across industries and within industries to understand the effects of the PPP on future reallocation as measured by future quits. Our analysis suggests that, in fact, additional utilization of PPP dollars was associated with excess quits several months after receiving PPP funds. In particular, we look at how much an additional week of PPP funding at time t is associated with a change in the level of quits between t and t+k periods, where k stands for months.⁵ Note that the main power of the empirical specification comes from the variation across industries in their utilization rates of PPP funding around the trough of employment in the pandemic, April through May 2020.⁶ We find that PPP funds equivalent to one additional week of payrolls might have increased quits somewhere between 2 percent to 5 percent over the following 12 months. This is the case after controlling for differences in industries and time periods when the quits rate is measured. Figure 4 plots the dynamics of this relationship with point estimates (orange) and the corresponding 90 percent confidence intervals (blue) and shows statistically significant effects later in the horizon. The average industry received PPP funds equivalent to 7.5 weeks of average payrolls in April 2020, and the largest disbursement was around 10.5 weeks. Our estimates suggest that compared to the average industry, the one that has received the largest amount would have experienced 6 percent to 15 percent higher quits over the following 12 months. This is an economically meaningful effect. For comparison, observe that throughout 2021, the average difference in quits between small firms versus large firms was about 15 percent (Figure 2).

Conclusion

The US labor markets experienced a relatively robust recovery over the past two years. We argue that the surge in quits is higher than what one would expect even in such a tight labor market.  Looking beyond the headline numbers for quits suggests that the aggregate pattern might be driven by small firms, and a purely cyclical story cannot explain currently high quits rates. We provide one potential explanation that relies on PPP loans and their utilization by small businesses during the pandemic and the following recovery and show that industries receiving larger PPP loans experienced elevated quits even 12 months after the PPP disbursement, suggesting that the availability of PPP funds might have prevented some “usual” reallocation from happening early on and thus subsequently created a pent-up demand for labor market reallocation. The COVID-19 pandemic introduced other important changes such as prevalence of remote work and shifting tastes for job-related amenities. More work is needed to explore whether these alternative mechanisms and changes could have anything to do with excess quits during this episode.