On June 12, 2019, EPI Senior Economist and Director of Policy Heidi Shierholz testified before the U.S. House of Representatives, Committee on Education and Labor, Subcommittee on Workforce Protections, regarding the Department of Labor’s March 2019 proposal to update the overtime salary threshold.
Chair Adams, Ranking Member Byrne, and members of the subcommittee, thank you for the opportunity to testify today. My name is Heidi Shierholz and I am a senior economist and the director of policy at the Economic Policy Institute (EPI) in Washington, D.C. EPI is a nonprofit, nonpartisan think tank created in 1986 to include the needs of low- and middle-wage workers in economic policy discussions. EPI conducts research and analysis on the economic status of working America, proposes public policies that protect and improve the economic conditions of low- and middle-wage workers, and assesses policies with respect to how well they further those goals. I was also the Chief Economist at the U.S. Department of Labor from mid-2014 to early 2017, where I worked extensively on the 2016 overtime rule.
My testimony first discusses the significance and purpose of overtime protections and provides background on overtime rulemaking. I show why the 2004 rule was fundamentally flawed and why the 2004 methodology is therefore an inappropriate methodology to use to set the overtime salary threshold. I show that the 2016 rule used a highly appropriate, albeit conservative, threshold. I also discuss the importance of automatic indexing of the threshold to ensure that it doesn’t erode over time. Finally, I describe the workers affected by the Department of Labor’s current proposal, including the impact on their earnings. Unless otherwise noted, the empirical analysis cited in this testimony is based on pooled 2016–2018 Current Population Survey Merged Outgoing Rotation Group microdata from the Bureau of Labor Statistics, following the same methodology used by the Department of Labor in its overtime analyses.
The significance and purpose of overtime protections
The Fair Labor Standards Act (FLSA) requires employers to pay their employees at least the federal minimum wage for all hours worked, and caps at 40 the number of hours an employee can work in a workweek without additional compensation, with employers being required to pay overtime premium pay of 1.5 times the employee’s regular rate of pay for all hours worked beyond 40 hours. These overtime protections ensure that employers have “skin in the game” when they ask employees to work long hours, which leads to two key outcomes: (1) employers are incentivized to hire more employees rather than overworking existing employees, and (2) employees are fairly compensated when they are required to work long hours. These effects make overtime protections a crucial part of the foundation of a vibrant middle class.
In the FLSA, Congress provided overtime protections to most workers, but directed the Secretary of Labor to exempt bona fide executive, administrative, and professional (EAP) employees from these protections. While most workers need overtime protections because they do not have sufficient individual bargaining power and control over their work to protect themselves from overwork for low pay, EAP workers could be exempted because they command enough bargaining power and responsibility to be able to successfully advocate for themselves. For an employee to be exempt from overtime protection based on the EAP exemption, they must satisfy each of the following three conditions: 1) they must earn a salary, i.e., not be paid by the hour; 2) they must pass the duties test, meaning that they are a bona fide executive, manager, supervisor, or highly trained professional based on the characteristics of the work they do; and 3) they must earn above the salary threshold.
The salary threshold can best be thought of as a rough proxy for a duties test—a bright dividing line that simplifies the determination of exemption for employers and employees alike, helps ensure that those with relatively low pay who do a great deal of nonexempt work are not taken advantage of, and reduces the misclassification of non-EAP workers as exempt. The duties test serves as a more specific mechanism for making the determination of exemption for salaried workers who earn above the threshold.
Background on the recent overtime rulemaking
Just three years ago, the Department finalized an overtime rule following an exhaustive, more-than-two-year rulemaking process.1 During this process, the Department met with over 200 organizations, including employees, employers, business associations, nonprofit organizations, employee advocates, unions, state and local government representatives, tribal representatives, and small businesses. The Department also received and reviewed over a quarter million public comments. In the 2016 Final Rule, the Department responded comprehensively to those comments and conducted a thorough economic impact analysis incorporating that input, along with a careful review of the academic literature.
The core provision of the 2016 rule was to increase the salary threshold under which most salaried workers are eligible for overtime pay when they work more than 40 hours per week from $455 per week ($23,660 for a full-year worker) to $913 per week ($47,476 for a full-year worker), the latter being the 40th percentile of the earnings of full-time salaried workers in the lowest-wage census region, which was at the time, and continues to be, the South. Further, the rule provided that the threshold would be updated every three years to the 40th percentile of the earnings of full-time salaried workers in the lowest-wage census region, in order that the threshold would not continually erode over time as overall wages rise. I project that the 2020 level of the threshold under the 2016 rule would be $982 ($51,064 for a full-year worker).
In November 2016—just before the 2016 rule was set to go into effect—a single district court judge in Texas enjoined the Department from enforcing the rule; the court later erroneously held the rule to be invalid. Instead of defending the rigorously determined threshold, the Department has proposed to rescind its 2016 rule and promulgate a new regulation with a much lower standard salary threshold. The core provision of the Department’s 2019 proposal is to set the salary threshold under which most salaried workers are eligible for overtime pay when they work more than 40 hours per week at $679 per week in 2020 ($35,308 for a full-year worker), which is the projection to January 2020 of the 20th percentile of the earnings of full-time salaried workers in the lowest-wage census region, currently the South, and/or in the retail industry, excluding nonexempt workers and workers who are not subject to the FLSA or who are not subject to the salary level test. The proposal does not include automatic updating. This methodology for setting the threshold is the same methodology used in the 2004 final rule.
The 2004 rule was deeply flawed
Unfortunately, the methodology for setting the standard salary threshold in the 2004 rule was fundamentally flawed. Prior to the 2004 rule, there were two sets of tests, each of which involved a duties test and a salary test. The duties test and salary test within each set had always worked together.2 One set of tests was the “long-test” set, which combined a low salary threshold with a stringent duties test (which included a 20 percent cap—40 percent in the retail industry—on the amount of time overtime-exempt employees could spend on nonexempt duties). The other was the “short-test” set, which combined a high salary threshold with a much more lenient duties test. Thus an employer who wanted to assert that a relatively low-paid employee was exempt had to show more rigorously that their duties were “bona fide executive, administrative, or professional” in nature, whereas for a more highly paid employee the employer did not have to make as rigorous of a showing.
In the 2004 rule, the Department included just one set of tests. For this set of tests, the Department created a “standard” duties test that was essentially the more lenient “short test” for duties. To remain consistent with the prior methodology, the Department should have paired this duties test with a higher salary threshold consistent with a short duties test. Instead, they used a low salary level consistent with a long duties test; in the 2004 final rule, the threshold is actually referred to as a long-test salary level.3 This was a fundamental error. Because of the mismatch between the duties test and the salary threshold in the 2004 rule, the methodology from the 2004 rule is not an appropriate methodology for setting the salary threshold. If the Department were to set the threshold according to the 2004 methodology in the final rule, it would be doubling down on a fundamentally flawed approach.
It is worth noting that because the Department erred in 2004 and paired a “short” duties test with a low, “long-test” salary threshold, the one and only way the 2004 methodology for calculating the standard salary threshold could be used appropriately would be if the Department were to strengthen the duties test to align with the historical long test and account for Congress’s intent that only bona fide executive, administrative, and professional employees be exempt from overtime pay. The Department notes, “Because the long [emphasis added] duties test included a limit on the amount of nonexempt work that could be performed, it could be paired with a low salary that excluded few employees performing EAP duties.”4 The Department could, for example, set a bright-line duties test requiring a large majority of a worker’s work to be exempt, as previous rules did.
The standard threshold in the 2016 rule was appropriate, albeit conservative
In 2016, the Department chose to correct the error in the 2004 final rule by retaining the same lenient duties test that was used in the 2004 rule, but correctly pairing it with a higher salary level consistent with a short duties test. In fact, the Department picked close to the lowest possible value that would be consistent with the historical level of the short test. According to the Kantor method, which was the method used to calculate salary thresholds from 1958 until the 2004 elimination of the long test, the Department should have set the value somewhere between $889 and $1,245.5 The threshold the Department chose, $913, was at the very low end of that range. The fact that the 2016 salary threshold was highly conservative by historic standards is underscored by two additional points. First, the 2016 rule covered far fewer workers than the threshold did historically. In 1975, more than 60 percent of full-time salaried workers earned below the threshold and so were automatically eligible for overtime based on their pay. By 2016, that share had dropped to less than 7 percent, and the 2016 rule would have only partially restored it, to around 33 percent. Second, if the 1975 rule had simply been updated for inflation, it would have been well over $50,000 in 2016, substantially higher than the $47,476 threshold in the 2016 rule.
Despite this clear record, the Department appears to have adopted two erroneous concerns of the district court judge who enjoined the 2016 rule: (1) that the 2016 salary threshold was too high because the number of newly overtime-eligible workers was large, and (2) that the large number of newly overtime-eligible workers under the 2016 threshold meant that the salary threshold was so high that it displaced the role of the duties test.6 Both of these concerns are deeply misguided. First, the raw number of workers affected by any increase in the salary threshold is an absurd metric to use to assess whether the threshold is appropriate, because the number is affected by factors that are wholly unrelated to the appropriateness of the threshold, including how long it has been since the prior update and whether the prior threshold was set at an appropriate level. The longer it has been since the previous update, the more workers will be affected, as inflation and the overall wage structure rise over time and erode the effective level of the threshold. In the case of the 2016 update, it had been over a decade since the prior update. And if the prior threshold had been set inappropriately low, as it was in the 2004 rule, the number of workers affected would need to be larger to correct the error.
The number of workers affected by any increase in the salary threshold can be decomposed into two main components: workers affected as a result of the passage of time since the previous update, and workers affected as a result of any change in methodology from the previous update. One way to isolate the “time” component is to simply look at how many workers would have been affected as a result of the 2016 rule if the methodology had been the same in 2004 and 2016—that is, if the 2004 threshold had been set at the 40th percentile of the weekly earnings of full-time salaried workers in the lowest-wage census region, as it was in 2016. In 2002—the data year the 2004 threshold is based on—the 40th percentile of the weekly earnings of full-time salaried workers in the lowest-wage census region, the South, was $660. If in 2016 the threshold had been increased from $660 to $913 instead of from $455 to $913, the number of workers affected would have been 2.9 million.7 That means that 2.9 million of the 4.1 million workers affected by the 2016 rule were affected as a result of the erosion of the effective level of the threshold since the prior update, not as a result of the change in methodology. The remaining 1.2 million workers were affected as a result of the change in methodology from the flawed 2004 methodology to the appropriate methodology of the 2016 final rule.
The other concern that the Department seems to adopt, that the large number of newly overtime-eligible workers under the 2016 threshold meant that salary threshold was so high that it displaced the role of the duties test, is flatly refuted by the Department’s own analysis. A figure in the 2016 final rule shows that 47 percent of white-collar workers who failed the duties test earned above the 2016 salary level.8 This means that of white-collar salaried workers who were eligible for overtime, nearly half—6.5 million workers—had their overtime-eligible status determined by the duties test alone, demonstrating that the duties test was not somehow “displaced” by the 2016 salary threshold but was, in fact, still essential to the exemption.9 Lowering the threshold in the interest of not displacing the duties test is solving a problem that the Department’s own analysis shows does not exist.
If the Department is concerned about the number of workers affected in a given year by raising the salary threshold to an appropriate level, it should simply phase in an appropriate threshold over time instead of reducing the threshold to an inappropriate level. For example, the Department could implement a three-stage phase-in, raising the threshold to $679 (or $35,308 for a full-year worker) in 2020; raising it to $866 (or $45,032 for a full-year worker) in 2021; and, finally, raising it in 2022 to be equal to the 40th percentile of the weekly earnings of full-time salaried workers in the lowest-wage census region. (I project that this will be $1,030—or $53,560 for a full-year worker—in 2021Q2; 2021Q2 data could be used to set the threshold for 2022).
A final point worth making about the 2016 rule is the scope of its economic impact. According to the Department, the 2016 rule would have led to an aggregate increase in payroll costs to business of roughly $1.5 billion dollars annually.10 That may sound like a large number—and the rule would certainly have a meaningful effect for workers who see higher pay, work fewer hours, or just gain clarity about their overtime-eligibility status—but total U.S. payroll costs are more than $8 trillion dollars per year.11 In other words, $1.5 billion is less than one-tenth of one percent of total U.S. payroll costs. The department does not need to promulgate a weaker rule than the 2016 rule in order to have a rule that does not have a disruptive effect on the labor market and the broader economy.
Automatic indexing is crucial
DOL has ample authority to index the salary threshold. The FLSA exemption from the minimum wage and overtime protections for EAP employees specifies that these exemptions must be “defined and delimited from time to time by regulations of the Secretary [of Labor].”12 Given that indexing is simply a means to ensure the threshold will remain current rather than continuously erode, DOL would be acting entirely reasonably and within its statutory authority13 to adopt indexing as a means to define and delimit the EAP exemptions in a timely manner.
Despite this authority, the Department is not proposing automatic indexing and instead says it is “committing to evaluate” the threshold more frequently going forward—in particular mentioning that it has an intention to go through notice and comment rulemaking every four years to update the rule using the same methodology as the most recent final rule, but noting that the Labor Secretary could forestall this process at his or her discretion.14
This is problematic on many levels. First, if the rule really is updated every four years through notice and comment rulemaking, that is still too long between updates and will leave workers behind in the meantime. Updating should occur no further apart than every three years, with one year being optimal, to reduce the degree of erosion between updates.
Further, increasing the frequency of rulemaking in this way is a terribly inefficient way for the government to operate. Rulemaking is extremely time- and resource-intensive, and it doesn’t make any sense to go through that process just to maintain the status quo. Automatic updating should be used to ensure that the standard doesn’t continually erode until such time that policymakers want to change the actual substance of a rule, at which point notice and comment rulemaking is appropriate.
Perhaps more importantly, vague commitments to update the rule, rather than implementation of automatic updating, is what has led to long periods of inaction in the past, with workers getting the short end of the stick as their protections erode over long stretches of time. My analysis, described in greater depth below, shows that if this rule is implemented as proposed in 2020, workers will earn $1.2 billion dollars less in 2020 than they would have earned under the 2016 final rule in the first year of implementation—but the annual losses will grow to $1.6 billion over the first 10 years of implementation due to the lack of automatic updating.
Finally, automatic updating is important for employers, too. With automatic updating, as opposed to a flexible intention to update the rule every four years, employers know exactly what to expect and when to expect it, which provides businesses the crucial predictability they need to plan for the future.
Millions of workers will be left behind by the Department’s current proposal
If the Department finalizes its new proposal, millions of workers who should get overtime protections will fall through the cracks. In its proposal, the Department estimates that 2.8 million fewer workers will be impacted under its proposal than under the 2016 rule.15 However, this is a vast underestimate, for two reasons. First, the Department uses pooled 2015–2017 data, benchmarked to 2017 wage and employment levels, and states that these figures “are the Department’s best approximation for impacts starting in 2020.”16 This leads to an underestimate because it doesn’t account for employment growth and other changes in the three years between 2017 and 2020. In my analysis, I correct for this issue by using more updated data—pooled 2016–2018 data, benchmarked to 2018 wage and employment levels—and inflating employment and wage levels based on Congressional Budget Office economic projections for 2018–2020.17
Second, the Department’s estimate of those left behind leaves out an entire group of workers who would be affected by the rule—those who will no longer get strengthened protections. To understand what a large omission this is, it is useful to keep in mind that there are two groups of workers who would be affected by any update to the overtime threshold. One group consists of those workers who get new protections under a new threshold—namely, salaried workers who pass the duties test who earn above the old threshold but below the new threshold. These workers are not legally entitled to overtime protections under the old threshold but would be overtime-eligible under the new threshold. Another large group of workers who are affected by any increase in the threshold are workers who get strengthened protections—salaried workers who earn above the old threshold and below the new threshold but who do not pass the duties test. These workers should have overtime protections under the old threshold—but because they earn a salary above the threshold, they are vulnerable to being misclassified by their employer as overtime-exempt. As mentioned above, research shows that this type of misclassification is pervasive. However, once the threshold rises above their earnings level, the status of these workers as overtime-eligible becomes very clear. In its estimate of how many workers would be left behind by its proposal, the Department ignores the millions of workers who will not get strengthened protections under its proposal but who would have gotten strengthened protections under the 2016 threshold.18
Our analysis shows that 8.2 million workers who would have benefited from the 2016 final rule will be left behind by this proposal. The 8.2 million workers left behind by this proposal are made up of 3.1 million workers who would have gotten new overtime protections under the 2016 rule and another 5.1 million workers who would have gotten strengthened overtime protections under that rule. The 8.2 million workers left behind include 4.2 million women, 3.0 million people of color, 4.7 million workers without a college degree, and 2.7 million parents of children under the age of 18. Of the 2.7 million parents left behind, more than half (1.4 million) are mothers. It’s worth noting that being left out of coverage under the FLSA’s overtime provisions has particular disadvantages for new mothers, since the Affordable Care Act’s protections requiring employers to provide time and space for nursing mothers to express milk at work apply only to employees who are not exempt from overtime pay.19
Further, because the Department is using a flawed methodology that sets the threshold far below where it was set using the appropriate methodology of the 2016 rule, workers will lose $1.2 billion dollars each year. This calculation includes both wages lost by workers who would have gotten new protections under the 2016 rule but would not get new protections under the current proposal, and wages lost by workers who would get new protections under both the current proposal and the 2016 rule but who would have gotten a larger raise under the 2016 threshold. The calculation does not include earnings losses by those who would have gotten strengthened protections under the 2016 rule but would not get them under the current proposal. As described above, the annual earnings losses would grow from $1.2 billion to $1.6 billion (in inflation-adjusted terms) over the first 10 years of implementation due to the fact that the current proposal does not include automatic indexing.
The Department’s current proposal lets down millions of workers and their families by failing to set the overtime salary threshold at a level that would lead to better use of one of the most precious resources of working families in this country—their time. The standard salary threshold in the proposal is so low that it fails to provide a true incentive for employers to balance the additional hours they ask of their workers with the costs of either overtime pay or of raising salaries to the new salary threshold. That incentive is inseparable from a fundamental principle embodied in the Fair Labor Standards Act—that workers should receive a fair day’s pay for a long day’s work.
The Department’s proposed rule—which at its heart is based on the notion that someone making $35,308 a year is a well-paid executive who doesn’t need or deserve overtime protections—flies in the face of those principles. As currently proposed, the Department’s proposal will have detrimental effects on workers, depart from decades of historical precedent, and undercut the purpose of the Fair Labor Standards Act’s overtime provisions, leaving behind millions of workers who would have been covered by the painstakingly determined 2016 rule. Congress should step in and pass the Restoring Overtime Pay Act, which codifies the 2016 rule, setting the threshold at an appropriate level and automatically updating it going forward, helping create a fairer economy while at the same time providing crucial predictability to employers and employees alike.
1. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32396 (May 23, 2016).
2. “The Department has always recognized that the salary level test works in tandem with the duties requirements to identify bona fide EAP employees and protect the overtime rights of nonexempt white collar workers” (Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32444 [May 23, 2016]).
3. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 69 Fed. Reg. 22121 (April 23, 2004); Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [request for information], 82 Fed. Reg. 34616 (July 26, 2017).
4. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32465 (May 23, 2016).
5. The Kantor method is a two-stage process: It involves first calculating the long-test salary threshold as the 10th percentile of earnings of exempt workers for low-wage regions, industries, and other low-wage groups, and then multiplying that threshold by somewhere between 130 and 182 percent—with the average multiplier being 149 percent—to arrive at the short-test salary level. The value of the long-test threshold for the 2016 rulemaking was $684. Following the two-stage process, to get to the short-test threshold, the Department should have multiplied $684 by somewhere between 1.3 and 1.82, which would have resulted in a short-test threshold range of between $889 and $1,245. See Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32463, 32467 (May 23, 2016).
6. The Department states: “The district court’s decision raised concerns regarding the large number of exempt workers—4.2 million—who earned between $455 and $913 per week and thus would “automatically become eligible” for overtime under the $913 per week standard salary level. The district court noted that this relatively high number indicated that the salary level was displacing the role of the duties test in determining exemption status. The Department acknowledges these concerns and, additionally, in this proposal seeks to update the standard salary level in a manner that does not unduly disrupt employers’ operations; dramatically shift employee salaries, hours, or morale; or result in adverse economic effects.” See Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [proposed rule], 84 Fed. Reg. 10909 (March 22, 2019).
7. Note that this is an underestimate to the extent that if the threshold were $660, there would likely be more people earning $660 instead of somewhat below $660 than what we actually see in the data, as some workers earning somewhat below $660 would have been bumped up to $660 so their employers could claim the exemption.
8. See Figure 3 of Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32465 (May 23, 2016).
9. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32465 (May 23, 2016).
10. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [final rule], 81 Fed. Reg. 32468-32469 (May 23, 2016).
11. U.S. Bureau of Economic Analysis, “Table 1.10. Gross Domestic Income by Type of Income.”
12. 29 U.S.C. § 213(a)(1).
13. See this letter from a group of law professors on DOL’s authority to index the salary threshold: https://www.regulations.gov/document?D=WHD-2015-0001-4585.
14. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [proposed rule], 84 Fed. Reg. 10914-10915 (March 22, 2019).
15. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [proposed rule], 84 Fed. Reg. 10951 (March 22, 2019).
16. Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [proposed rule], 84 Fed. Reg. 10950 (March 22, 2019).
17. Heidi Shierholz, More Than Eight Million Workers Will Be Left Behind by the Trump Overtime Proposal, Economic Policy Institute, April 2019.
18. While the Department does not include this group in its analysis of how many fewer workers will be impacted under its proposal than under the 2016 rule, it does mention them elsewhere in the proposal: “The Department also anticipates that 3.6 million employees paid between $455 and $679 per week who fail the standard duties test (i.e., that are and will remain nonexempt)—2.0 million salaried white collar workers and 1.6 million salaried blue collar workers—will have their nonexempt status made clearer because their salary will fall below the proposed threshold” (Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees [proposed rule], 84 Fed. Reg. 10911 [March 22, 2019]).
19. Heidi Shierholz, “Millions of Working Women of Childbearing Age Are Not Included in Protections for Nursing Mothers,” Working Economics Blog (Economic Policy Institute), December 10, 2018.