Limiting abusive employment practices requires new data
Understanding who does what at work is critical for effective overtime, worker classification policy
Key questions in labor policy turn on information about the content and structure of jobs that cannot be answered well using existing data. As a result, it is difficult to design and understand the implications of potential changes to rules that determine overtime coverage and appropriate employment classification. Collecting additional information about people’s activities at work by occasionally adding supplemental questions through the Current Population Survey would fill critical knowledge gaps about basic facts, generate insights into the evolving organization of work that would be of broad interest to researchers, and enable better-informed policymaking.
To keep up with evolving labor market practices, policymakers need data
The nature and structure of work are constantly changing. New industries and occupations emerge while those that have been left behind by the evolution of the economy fade away. The prevalence of arrangements like outsourcing, the use of independent contractors, and remote work rises and falls. Failure of policy to keep up with current practices creates opportunities for exploitation, such as wage theft and worker misclassification.
Despite sometimes being constrained by the limited availability of enforcement resources and focusing the resources it does have on relatively low-paying industries, the Wage and Hour Division of the Department of Labor (DOL) routinely collects well over $100 million per year in back pay for workers who have been improperly denied overtime compensation. Data on the determinants of worker classification are so limited that it is difficult to say how common it is for workers to be inappropriately classified as independent contractors, a misclassification that can cost them legal protections and eligibility for benefits. Unfortunately, policymakers face a significant obstacle to preventing these abuses: there are essentially no individual-level data available on who does what at work and how jobs are structured to guide policy development.
Overtime policy illustrates the need for more detailed data
In early September, DOL published a proposed rule that would raise the earnings threshold below which all salaried workers are entitled to premium pay for hours worked beyond 40 in a week.1 A salary threshold that is “too low” leaves low-paid workers vulnerable to abuses such as being given a salary and a nominally supervisory title, being exempt from overtime coverage on that basis, and then being asked to spend long, uncompensated hours doing work that should be subject to overtime protection, like stocking shelves.2 The move makes the Biden administration the fourth consecutive administration to at least attempt to change overtime rules. Each of those efforts has focused on changing the salary threshold, long used as a tool for simplifying overtime converge determinations for lower-paid workers who are unlikely to be doing exempt work, rather than trying to change the relationship between workers’ duties and their overtime eligibility.
The proposed Biden rule is in line with the proposal issued by the Obama administration, given the eight years that have passed, setting the salary threshold somewhat higher in nominal terms ($1,059 per week/$55,068 annually vs. $913 per week/$47,476 annually under the Obama proposal) but lower in the earnings distribution (the 35th percentile of weekly earnings for full-time salaried workers in the South, the lowest-wage Census region vs. the 40th percentile under the Obama proposal). The Biden proposal would also automatically update the salary threshold every three years based on the same 35th percentile rule used to set it in the first instance.
The Obama proposal, however, never took effect because a federal district court judge in Texas enjoined it. The judge claimed that the rule relied too heavily on workers’ salaries to determine overtime eligibility, contrary to the statutory requirement that these determinations be based on workers’ duties (those whose primary duties are executive, administrative, or professional are exempt from overtime coverage). In other words, too many workers would be made eligible for overtime coverage by virtue of their relatively low pay even though an assessment of their duties alone would lead to them being exempt from overtime coverage.3
However, the number of workers who would gain overtime coverage due to any particular change in the salary threshold (i.e., the number with earnings below the threshold whose duties would not have made them eligible) is subject to a great deal of uncertainty. In order to estimate the effects of a proposed rule, the Department of Labor needs to know how many workers are already covered by virtue of their duties at various earnings levels. Unfortunately, the best available information about overtime coverage has important limitations.
In each overtime rulemaking process since the George W. Bush administration, DOL has analyzed the effects of the proposed rules using estimated probabilities that workers within each occupation would be exempt from overtime coverage that were originally produced by Wage and Hour Division staff in 1998. These probabilities are reported in very wide ranges (e.g., 10% to 50% or 50% to 90%) for many occupations, suggesting substantial scope for workers’ duties to vary. The probability estimates do not address how that within-occupation variation in duties is related to earnings.4 There is some information about how job duties vary across occupations from O*NET, but it does not address the within-occupation relationship between duties and earnings, either. Finally, the probability estimates are now 25 years old, and a lot has changed since they were originally produced.
Regular collection of data on the work activities and earnings of individual workers could address these limitations and help policymakers set a salary threshold that efficiently guarantees overtime coverage to workers who should be covered. Such data could help illuminate any tradeoff between simplifying overtime eligibility determinations/enforcement of overtime rules and making the correct overtime coverage determination for workers at the margin. Unfortunately, these data do not currently exist.
We have the ability to collect this kind of information
While careful survey design would be critical, the infrastructure to collect this kind of information exists and is used regularly by a variety of agencies. Every month, the Census Bureau conducts the Current Population Survey (CPS), which provides the data underlying indicators like the unemployment rate, the labor force participation rate, and the employment-population ratio. The survey consistently collects information on labor market experiences, earnings, and demographics, and regularly includes supplemental questions on topics of interest to sponsoring agencies. Previously conducted CPS supplements cover topics ranging from voting to tobacco use to fertility to income. DOL, a natural potential sponsor for a survey on work activities, sponsored a supplement focused on unemployment insurance filing in 2005 and 2018. The length of these supplements ranges from about a dozen questions or so to lengthy questionnaires that take up over 100 pages.
A Work Activities Supplement to the CPS would result in a publicly available, individual-level dataset that includes information on earnings, work activities, and demographics for a nationally representative sample consisting of more than 10,000 people. This would allow for flexible analysis of what people do at work, how their jobs are structured, and how those things relate to their other characteristics. While more ground could be covered in a longer survey, even 15 questions or so could provide insights that would be extremely valuable for policy. Questions could be tailored to inform policymakers’ thinking not only on overtime rules but also on topics like employment classification, which similarly hinges on information not currently collected in any representative way. Conducting a Work Activities Supplement every two to three years would probably be often enough to avoid falling too far behind changes in how the content of jobs and the organization of work are evolving. Regular data collection would be especially helpful if the overtime salary threshold is to be updated automatically every three years, as the Biden administration has proposed.
The cost of a Work Activities Supplement is likely to be small, especially compared to the value of the economic activity affected by the regulations it could inform. The Contingent Work Supplement, which also adds detailed questions about people’s work arrangements to the CPS, was last conducted in July 2023. The agency’s budget request for fiscal year 2023 does not mention a specific amount associated with that survey, but in 2017, the previous time it was conducted, BLS requested about $1.6 million to conduct the survey. Even if costs have risen since then or additional funds are required to develop a survey covering entirely new topics, the cost of a Work Activities Supplement is likely to remain on the order of a few million dollars. In comparison, DOL’s analysis of the Biden administration’s proposed overtime rule indicates that it would transfer about $1.3 billion per year from employers to employees, and it would have direct costs to employers associated with familiarization, adjustment, and managerial costs of about $660 million per year. A Work Activities Supplement could conceivably “pay for itself” by reducing employer costs associated with one related regulation by one percent (or less), setting aside all its other benefits.
Issues related to overtime and employment classification policy are ripe for this kind of investigation
The core of a Work Activities Supplement could initially include questions related to (1) the performance of the executive, administrative, and professional duties relevant to exemptions from overtime coverage, and (2) the nature of the relationship between workers and those who hire them. Covering key elements of these topics would balance the potential to improve our understanding of the labor market with value to policy development. Doing so would also require only a modest number of questions, keeping respondent burden light.
Table 1 lists the key facts that are important to understand about a person’s work to determine whether they are (1) exempt from overtime coverage based on executive, administrative, or professional duties, and (2) appropriately classified as an employee or an independent contractor.5 Developing survey questions to cover these topics requires expertise and is beyond the scope of this memo, but this table suggests that among 15 questions focused on these core policy areas, ten questions could be used to cover activities related to overtime exemption, and five could cover employment classification. If the process of developing questions reveals that more than 15 are needed, there would be room to expand the supplement modestly without overburdening respondents.
A somewhat longer Work Activities Supplement would provide an opportunity to understand a variety of related topics beyond these core policy issues. There are many emerging questions related to remote work, including questions about how it may interact with overtime and employment classification policy. Including questions about the role of technology in jobs and the manual, cognitive, or routine nature of work would enhance researchers’ ability to understand the continuing evolution of these trends by connecting this information to other relevant personal characteristics of workers.
Important research over the last 20 years or so has considered the role of technological change in the evolution of income inequality, as well as the polarization of the labor market around the content of work in different jobs (e.g., Card and DiNardo, 2002; Autor, Levy, and Murnane, 2003; Autor, Katz, and Kearney, 2008; Autor and Dorn, 2013; Beaudry, Green, and Sand, 2013; Hunt and Nunn, 2019). In both cases, existing data make it easiest to analyze occupations, even though exposure to changes in technology and the actual duties associated with one’s job are specific to individuals.
Policy benefits flow from understanding the labor market better
Fielding a Work Activities Supplement to the CPS would fill important gaps in our understanding of the labor market. Conducting such a survey once would provide basic facts about pay and the content of work that are currently unknown. Conducting it repeatedly would help track the evolution of jobs over time. This new data collection would both advance our understanding of how jobs evolve over time and improve the ability of policymakers to make informed decisions regarding regulations that impact millions of workers. Adding detailed information about work activities to the data already available in public-use CPS microdata would also deepen our understanding of recent labor market trends and open up new lines of research. The knowledge gained by collecting and analyzing these data would create opportunities for improved policymaking in areas where policymakers are currently limited to taking shots in the dark.
Some salaried workers are not required to be paid a premium wage for hours worked beyond 40 in a week (i.e., they are “exempt” from overtime coverage) under the Fair Labor Standards Act (FLSA). In order to be exempt, workers must “pass” the salary level test (by earning a salary that exceeds some amount set by regulation, currently $684 per week or $35,568 annually) and the duties test, by having primary duties that fall within one of several exemptions established by the law. The most common exemptions are those related to executive, administrative, and professional (EAP) duties.
Conversely, a salary threshold that is “too high” could potentially lead firms to adopt inefficient personnel policies in order to avoid paying the overtime premium or make it harder for some workers to invest in their careers by working long hours.
It should be noted that this reasoning was hotly contested at the time and in the following years. Some argue that workers earning below the salary threshold cannot be bona fide executive, administrative, or professional workers because they lack the status and ability to protect themselves from exploitation conferred by higher pay. However, barring a sudden and dramatic shift in the composition of the federal judiciary, this reasoning will likely influence future litigation around overtime regulations.
While DOL understands that higher-earning workers within an occupation are more likely to be exempt from overtime coverage than lower-earning workers and has attempted to model that relationship, there is no way to know how closely its approach reflects reality.
Under the FLSA, whether a worker should be classified as an employee or an independent contractor is determined using the “economic realities test,” which is intended to take a holistic approach to determining whether the worker is dependent on the business that has hired them (i.e., is an employee) or is engaged in a business of their own. This test does not have any specific requirements but is based on the totality of the relationship between the worker and the hiring entity. California has adopted its own standard for worker classification called the “ABC test,” and many Democrats have endorsed legislation that would adopt a similar standard in federal law. This test, while dealing with similar elements of the relationship between workers and hiring entities, relies on a few bright-line rules to establish independent contractor status.