This post was co-authored by Daniel Sternberg and Sarah Gustafson.
The federal government has made nearly $700 billion available to small businesses through the Paycheck Protection Program (PPP)—a forgivable loan initiative designed to help small businesses keep their employees on payroll as they deal with COVID-19-related disruptions to their work and revenue streams.
Run payroll and benefits with Gusto
As we have detailed in previous reports, the necessary shelter-in-place mandates across the US have resulted in millions of small businesses being forced to make extremely difficult decisions in order to survive, including unprecedented levels of layoffs, furloughs, reduction in employee hours, moving to new states and creating completely new businesses. Aid from the PPP program was designed to be a crucial lifeline to buy enough time for the US to reopen for business.
To determine the initial effectiveness of government relief efforts, Gusto analyzed data from nearly 27,000 of our small business customers who reported receiving PPP loans and compared it to platform data from our 100,000-plus small business customers nationwide. The report below shows that PPP aid has helped to provide stabilization from the initial free fall in March ‘20, with strong increases in hiring and rehiring beginning in the second half of April ‘20.
But our report also shows that PPP aid has not yet been enough to create a return to pre-COVID employment levels. New legislation passed by Congress extends the timeline and eases restrictions on how funds are used, which may help to speed up initial recovery efforts. Even with these changes, small businesses remain in a race against the clock to set up shop, rehire employees, and take care of fixed costs beyond payroll. And they still must navigate varying timelines for full reopening. In addition, most of the businesses that received PPP funding will have spent their PPP funds within the next few weeks while many of these businesses are still only able to partially operate.
Gusto will continue to monitor platform data and additional sources to better understand the effectiveness of PPP and other forms of government aid.
- Hiring rates are ramping up: Hiring and rehiring rates in late April through early June were nearly twice as high for companies that reported receiving PPP loans compared to those that did not (34.2% versus 18.4% hired at least one employee during this period).
- Bridging the gap: Federal loan programs helped small businesses with their April and May hiring efforts, but not enough to make up for the deep job cuts they made in March. Businesses that reported receiving a PPP loan were more than twice as likely (+139%) to rehire at least one employee from the last week of April through mid-May than businesses that did not. However, these companies were also still 65% more likely to terminate at least one employee during that same period.
- The first cut is the deepest: Businesses that received PPP loans were those that made the deepest job cuts before the CARES Act passed in late March ‘20. Despite higher rehiring rates for employees that were terminated between the onset of the COVID crisis and the beginning of PPP loan payouts, the percentage of employees that were out of work is still higher for PPP-receiving companies (86.9% of employees were still employed as of mid-May versus 89.6% for companies that did not report receiving a PPP loan).
Impact on PPP recipients
PPP-receiving companies were those that were the hardest hit in the final weeks of March.
The most likely explanation for this trend is that companies that needed the support the most were also those most likely to apply for a loan. Small businesses that received PPP funding had a 30% higher termination rate on average in the peak week than those that did not report receiving funding (3.5% versus 2.7% in the week of March 23).
Figure 1 below shows the week-by-week termination rates across all Gusto small business customers starting before shelter-in-place orders were issued. For example, if there were 100 workers employed across all customers the week of February 3, and 10 of those employees were terminated in the following week, the Termination Rate would be 10% for the week of February 10. If two more employees were terminated the following week, the Termination Rate for the week of February 17 would be 2.1% (where the denominator is the remaining 90 employees from the prior week).
Figure 1: Week-by-week Termination Rates for companies that did or did not report receiving a PPP loan
Hiring rates in April and May were higher for companies that received PPP loans
Hiring sharply declined for both groups in March, but the companies that reported receiving a PPP loan had higher hiring and rehiring rates in late April through mid-May compared to companies that did not report receiving a loan. The hiring rate for PPP-receiving companies was 34% higher for the week of May 11 than for companies which did not report receiving a loan (2.4% versus 1.8%). Hiring rates across the two groups approached the same rate by the beginning of June.
Similarly, the rehiring rate, which relates to the subset of all hires that were rehired by the same company from which they had been terminated, for PPP-receiving companies was 81% higher than for other companies (0.65% versus 0.36% for the week of May 11). Rehiring remained higher for the PPP receiving companies into the first week of June.
Figures 2 and 3: Week-by-week Hiring Rate and Rehiring Rates for companies that did or did not report receiving a PPP loan
The combination of termination data and hiring data is shown in Figure 4 below as the Net % Change in Headcount. For example, if there were 100 workers employed across all customers the week of February 3, and, in the following week, 10 of those employees were terminated and five new employees were hired, the Net % Change in Headcount would be -5% for the week of February 10. The subsequent week’s denominator would be the 95 employees that were employed the week of February 10.
Figure 4 shows that both the deep cuts in March and the steep climb in hiring in April and May are more extreme for companies receiving PPP loans. By the first week of June, both groups saw equal changes in net headcount, which indicates that the full impact of PPP loans for many companies has already played out.
Figure 4: Week-by-week Net % Change in Headcount for the weeks of Feb 3–June 1 for companies that did or did not report receiving a PPP loan
Companies that reported receiving a PPP Loan were more than twice as likely (+139%) to rehire at least one employee from the last week of April through early June than companies that did not provide PPP loan data. However, these companies were also still 68% more likely to terminate at least one employee during this period (17.5% terminated at least one employee versus 10.4% for companies that did not provide PPP loan data, as shown in Table 1 below). It’s evident that while these companies are deriving some support from federal loan programs for the purposes of hiring, the employment landscape for small businesses is still especially fraught.
Table 1a, 1b: Employer Data — Termination and rehiring actions taken by companies with PPP loans versus those for which we have no PPP data
Table 1a. March 16–April 30
|Reported Receiving PPP||% Terminated at Least 1 EE||% Hired at Least 1 EE||% Rehired at Least 1 EE|
Table 1b. May 1–June 7
|Reported Receiving PPP||% Terminated at Least 1 EE||% Hired at Least 1 EE||% Rehired at Least 1 EE|
Table 2 below presents a similar story to Table 1 but looks at the employee-level experience instead of the actions taken by employers. Table 2 shows the percent of all employees that had been employed prior to March 16 and were terminated by April 30, the percent of those terminated employees that were subsequently rehired by June 7, and the resulting net employment rate for that set of employees. For example, if there were 100 employees working the week of March 9, and 10 of those were terminated before PPP loans were disbursed, the termination rate would be 10%. If five of those terminated employees were rehired, the rehiring rate shown in the third column below would be 50%. The net employment rate would be 95% in the final column.
Table 2: Employee Data — Termination and rehiring patterns for employees working for companies with PPP loans versus those for which we have no PPP data
|EE Belongs to a Company that Reported Receiving a PPP Loan?||% EEs Terminated Pre-PPP (March 16 – April 30)||% of those Terminated EEs that were rehired after PPP (May 1 – June 7)||% of beginning EEs still employed by June 7|
|We don’t know||12.4%||10.1%||88.9%|
Employees terminated in March who worked for companies receiving PPP Loans were 24% more likely to be rehired in May through early June than those working for companies for which we have no PPP data. Table 2 above shows that 12.5% of employees that work for companies that reported a PPP loan and were terminated in the last two weeks of March were rehired between May 1 and June 7, versus 10.1% for companies that did not report receiving a loan.
However, despite higher rehiring rates for employees that were terminated between the onset of the COVID crisis and the beginning of PPP loan payouts, the percent of employees that were out of work was still higher for PPP-receiving companies (86.1% of employees were still employed as of early June versus 88.9% for companies that did not report receiving a PPP loan). It appears that PPP loans to date have not made up for the deep cuts that were made in March for this subset of companies.
Figure 5 below shows the cumulative percent change in headcount for the two company categories as compared to employment in the first week of March. The companies receiving PPP loans, shown in orange, have not quite caught up to the non-PPP companies by the last fully observed week.
Figure 5: Cumulative % Change in Headcount since the first week of March for companies that did or did not report receiving a PPP loan
Sam Gilbert of Temescal Brewing in Oakland, California always knew he wanted to build a gathering place for his community. But even more than that, Sam wanted to build a company that was a community unto itself—that created good jobs with good benefits for those on his staff. “Of all the things that are part of our mission and values, what comes first isn’t beer quality, or even our customers, it’s our staff,” Sam says. “We exist to serve our employees, not the other way around.”
Earlier this year, Temescal Brewing was bursting at the seams with customers, with plans underway to build an expanded manufacturing facility in West Oakland. But shelter-in-place orders earlier in March forced Temescal Brewing to make the difficult decision to lay off the majority of its staff as it waited for the crisis to pass and federal relief to come through. The company had every intention of hiring back staffers when business resumed and has pivoted to an online delivery model in the meantime. Temescal Brewery recently received PPP aid and is taking a phased approach to rehiring while offering no-contact curbside delivery from its brewhouse.
Impact by industry
For companies that received PPP loans, Net % Change in headcount was significantly lower in March and subsequently higher in May in both Retail and Food & Beverage than other industries, although headcount for Food & Beverage companies was much more severely impacted than Retail. The steep drops shown in Figure 6 below for the weeks of March 16 and March 23 indicate substantial cuts to headcount were worse for the highlighted industries than the average of all other industries. Food and Beverage companies that reported receiving PPP loans saw a single-week headcount drop of -8.8% the week of March 16, whereas Retail companies that reported receiving PPP loans only dropped by -4.1%, and companies in other industries with PPP loan data reached their low point of -2.1% reduction in headcount the week of March 23.
The data corresponding to the week of May 11–17 shows that, in the wake of PPP loan approval, hiring slightly outweighed terminations more for companies that received PPP in the Retail and Food & Beverage industries than for all other industries. There appears to be evidence that these companies, all of which reported receiving PPP loans, were able to bring back employees beginning in late-April as loans were paid out. The net percent change in headcount for the specified week of May reached +2.2 and +2.1% for Food & Beverage and Retail companies, respectively, and only +1.3% for all other industries.
By early June, the variation in net hiring across these highlighted sectors and all other industries had effectively disappeared.
Figure 6: For companies that reported receiving a PPP loan, Net % Change in headcount by industry
Unfortunately, the cumulative change in headcount from the first week of March through early June shows that companies in the Retail and Food & Beverage industries suffered a more persistent drop in headcount than other industries. While Figure 6 above shows that companies receiving PPP loans in these industries had stronger hiring than other industries in May and into June, Figure 7 below indicates that it simply was not enough to return these companies to pre-COVID employment. Food & Beverage companies were the hardest hit and have not come anywhere close to employing the same number of employees they did prior to COVID-19’s onset. These companies bottomed nearly -19% of their pre-COVID headcount in the end of April, and they have only rebounded slightly to -11% by early June. Retail companies have recovered to -1% of their pre-COVID headcount from a low of -10%. Companies that received PPP loans and were not in Food or Retail are now even with their pre-COVID headcount after hitting a low of -5% in April.
Figure 7: For companies that reported receiving a PPP loan, Cumulative % Change in headcount by industry since the first week of March ‘20
Net Hiring Rate in May was higher for companies that received PPP loans, even when controlling for Industry and Location.
To further isolate the relationship between PPP loans and employer hiring and termination behavior, we built a model that controlled for the impact of both industry and location (state as well as metropolitan statistical area). This model showed that, when controlling for industry and location, companies that reported receiving a PPP loan had an 84% higher Net Percent Change in Headcount from the first week of March to mid-May. The baseline Net % Change in headcount in our model was 1.79%, and those companies which received PPP loans clocked in at 3.29% (1.5% higher in absolute terms).
Across all industries, companies appear to be receiving the maximum loan amount stipulated by the CARES Act (the equivalent of 10 weeks of payroll). Table 4 below shows the median loan amount per employee within each industry. Loans in Arts & Entertainment and Technology were much higher than other industries, such as Food & Beverage (where the loan amount per employee was less than half of that in Tourism). This is reflective of the typical wages of employees across sectors, which would have dictated the size of the Paycheck Protection loan for which the employer applied.
Table 4: Median loan amount per employee, broken out by industry
|Industry||Median PPP Loan Amount Per Employee ($)|
|Food & Beverage||$4,224|
|Salon & Spa||$6,302|
|Sports & Recreation||$3,973|
Gusto is a modern, online people platform that helps small businesses take care of their teams. In addition to full-service payroll, Gusto offers health insurance, 401(k)s, compliance and expert HR, and more. The company serves over 100,000 businesses nationwide and has offices in Denver, New York City, and San Francisco.
The analyses described in this report were based on four separate datasets obtained from Gusto’s small business payroll, benefits, and HR applications. The first dataset used payroll data to capture employee hours, wages, and terminations (Employee Level Employment Data). The second dataset used data from employers’ actions with respect to hiring and terminating employees on a weekly basis (Employer Level Hiring and Termination). The final two datasets contained PPP information from Gusto customers who replied to our PPP Survey or who are tracking their progress to PPP loan forgiveness using Gusto’s payroll tools.
Our flag for whether a company received a PPP loan was gathered from two sources: 1) an optional survey sent to company administrators and 2) an in-product tool that companies may elect to use in order to track their progress towards earning PPP loan forgiveness. Given that both of these data sources are opt-in, we acknowledge we do not have perfect separation between our comparison groups in the above report. However, given the number of responses we’ve measured and the distinct trends in responses for the PPP group, we believe that improved separation would only increase the gap in behavior for these groups of companies. The findings in this report could only be inverted if the population that actually did receive a PPP loan but had not informed Gusto through the survey or the forgiveness tracking tool had an extremely distinct profile or hiring/termination behavior than those which have provided their PPP information.
Hiring and Termination Data
A given employee on Gusto can have multiple “employments,” since an employer can potentially hire, terminate, and rehire the same employee multiple times. In our Hiring & Termination dataset, a hire corresponded to an employer creating a new employment with a hiring date, a rehire corresponded to an employer creating a new employment with a hiring date for a previously terminated employee, and a termination corresponded to the entry of a termination date for a given employment. Layoffs corresponded to terminations where the employer listed the reason for the termination as a “layoff” (one of the choices from a standardized list in Gusto’s termination flow). In order to capture a more time-sensitive view of employer activities, we recorded hires, terminations, and layoffs on the date that they were entered into the system, rather than their effective date. Terminations are also coded as voluntary or involuntary.
Employee hires, layoffs, and terminations were aggregated weekly for a given company. Hiring, termination, and layoff rates reported in this document are based on the number of times each event occurred in a given week or month, divided by the number of employees who were employed at the beginning of the previous week.
As furloughing is not data that we explicitly collect from employers, this category of employment status is not being measured on a weekly basis. Gusto’s approach uses assumptions based on typical hours worked for a given employee in previous months to determine, on a monthly basis, whether an employee has been furloughed. Given the lack of granularity for this measurement, furloughing is excluded from this particular report’s analysis. It should be noted that the analyses around terminations and rehiring, therefore, do not include employees who had been temporarily furloughed.
Locations reported in this report are based on the filing location of the employer. Industries reported in this document are based on self-report from customers within Gusto’s product.
Mixed Effects Model
To better isolate the impact of PPP loans on Net Percent Change in Headcount, we built a Mixed Effects model that included the Number of Employees (mean deviated) and a flag for whether the company received a PPP loan as fixed effects, and location and industry as crossed random effects. Effectively this model controls for location (both state and MSA) and industry to try and better understand the impact of the PPP loan.
While our overall model didn’t explain a meaningful amount of the variation in Net Percent Change in Headcount, the results indicated with 95% confidence that companies that reported receiving a PPP loan saw a Net Percent Change in Headcount 1.0% to 1.6% higher (absolute) in the week of May 11 versus companies that didn’t report receiving a PPP loan. The impact of the PPP loan is, if anything, likely being underestimated by our analysis. The group of companies that are not flagged as “Received PPP” likely contains companies that applied and received funds and simply did not elect to inform Gusto.
To compare point-in-time Net Hiring Rate results across industries, we ran several pairwise hypothesis tests—two-sample Kolmogorov-Smirnov (KS) Tests, as well as two-sample t-tests—comparing individual industries against all other industries. The null hypothesis of the KS test is that both groups being compared have been sampled from populations with identical distributions (e.g. that both populations have the same medians, variances, or means, but does not compare any one summary statistic in particular). The null hypothesis of the two-sample t-test is that the true difference between the sample means is 0.
Only industries which yielded significant results were highlighted in this report (Food & Beverage, Retail).
 Data collected from survey responses and from our opt-in product solution for tracking qualification progress to loan forgiveness. Our two groups for comparison are 1) Received PPP Loan and 2) Unknown PPP status. We are confident that the companies in the PPP group (1) received loans. There are certainly some companies (an unknown number) in the Unknown group who received PPP loans and simply did not report it to Gusto. A discussion on this is given in the Appendix.
 This analysis uses a point-in-time dividing line of April 30 to approximate pre-PPP behavior and post-PPP behavior across all companies, and does not adjust the timing depending on the date when a company received their PPP loan. April 30 was selected because, at that point in time, close to half of Gusto’s PPP recipients had received their funding.
 Point-in-time hypothesis tests comparing Net Hiring Rate for companies receiving PPP loans across industries for the weeks of March 16 and also May 11 yielded statistically significant results. Industries other than Retail and Food & Beverage did not significantly differ from the overall observed data.
 A highlighted list of Industries, selected from a longer list of industries that had at least 100 respondents to Gusto’s survey and tracking tool.