New business practices and communications technologies have led many workers to supply their labor outside of the traditional employment relationship (so-called W-2 employment). Yet despite the growth of the “gig economy,” there is surprisingly little official data that tracks the extent to which alternative work arrangements have increased or decreased in recent years. This lack of data is not only problematic for scholars and industry leaders, but also for policymakers and regulators who must modernize our nation’s labor policies.
Fortunately, the first official survey of contingent and alternative work arrangements in more than a decade was released today by the Bureau of Labor Statistics (BLS). How has the labor market changed?
According to the newly released data, an estimated 15.5 million U.S. workers have alternative arrangements for their primary employment—this includes independent contractors, on-call workers, temporary help agency workers, and workers provided by contract firms. Notably, this does not include workers who have a traditional main job, but engage in some alternative work on the side (e.g., a Lyft driver who works occasionally on weekends but has a fulltime job during the week).
The number of workers with alternative employment arrangements for their main job is up from 12.1 million workers in 1995 and from 14.8 million in 2005. Figure 1 shows that alternative arrangements continue to be an important portion of the labor force. However, adjusting for growth in overall employment, the fraction of workers in alternative arrangements has actually fallen slightly since 2005, from 10.7 to 10.1 percent.
This surprising small decline in the share of alternative employment in primary jobs is at odds with some other research. For example, Katz and Krueger (2016) find an increase in alternative work between 2005 and 2015. It may be that the BLS focus on primary employment is obscuring a broader increase in nonstandard employment. Economists Anat Bracha and Mary Burke find much higher rates of informal work—estimating that a third of adults engaged in some form of nonstandard work—when they focus on both primary and secondary employment, using a separate Federal Reserve survey conducted in 2015 and 2016.
However, the rate of multiple jobholding appears to have declined in recent decades, leaving a puzzle that remains to be explored. A possible explanation is that informal workers are less likely to report employment in surveys, or do not realize that they are classified as independent contractors. Conversely, the relatively constant share of alternative employment in primary jobs is consistent with research examining administrative tax data that show workers with self-employment earnings from schedule C profits growing from 10 percent to 10.7 percent of the workforce from 2006-2014.
The BLS found that specific types of alternative arrangements have evolved somewhat differently since 1995, as shown in figure 2. Temporary help agency workers and workers provided by contract firms remained at roughly their 2005 shares of the labor force.1 On-call workers declined slightly, while the share of independent contractors declined about 0.5 percentage points. Independent contractors include many “gig” workers, but the category is much broader, including free-lance writers and consultants, among others. Increases in gig employment may have been masked by offsetting declines for other workers, though we will have to wait until September to answer this question.
The BLS also examines what it calls contingent work, a distinct categorization defined by workers’ tenures and expectations of the longevity of their current working relationship. Using the most expansive definition2, the fraction of the labor force engaged in contingent work has actually fallen sharply from 4.9 percent in 1995 to 3.8 percent in 2017 (not shown).
This decrease in contingent and alternative work seems to be at odds with the rise of the online gig economy, where platforms like Uber and Taskrabbit connect workers to customers. One possibility is that today’s BLS estimates are not fully capturing gig workers: if gig workers expect their jobs to last, they would not be considered to engage in contingent work. Today’s BLS data does not allow for separate identification of gig workers, though the BLS will subsequently release its first-ever estimates of the share of the labor force whose primary employment is from mobile apps or websites. The best available estimate, provided by economists Katz and Krueger, is that only about 0.5 percent of workers are engaged in such work as a primary job.
Implications for policy
Armed with up-to-date, accurate data, policymakers and regulators can work to keep regulations relevant and appropriate to the modern labor market. One particularly pressing policy issue is the classification of contingent workers: should participants in the “gig economy,” and alternative work arrangements more broadly be treated as employees or independent contractors? As these new data demonstrate, a sizable share of workers in the United States remain outside the traditional employment structure and consequently lack many of the protections and benefits that come with being a traditional employee. Economists Jackson, Looney, and Ramnath document many of these disadvantages, also finding that the Affordable Care Act played an important role in providing health insurance for many workers in alternative arrangements.
A 2015 Hamilton Project policy proposal authored by Seth Harris (Cornell University) and Alan Krueger (Princeton University) presented a set of rules for a new, intermediate classification of employment: the independent worker. This classification bridges the gap between traditional W-2 employees and contractors, acknowledging unique characteristics of independent workers (some but not all of whom work in the gig economy). For example, independent workers choose their own hours of work—as do contractors—but they typically maintain an ongoing relationship with a platform (e.g., TaskRabbit), as do W-2 employees. This group would be a subset of those with alternative work arrangements, appropriate for workers who have relationships with intermediaries like TaskRabbit but retain control over their own work hours.
In their proposals, Harris and Krueger propose that independent workers have the right to collectively bargain as well as protection from various forms of employment discrimination. However, independent workers would not qualify for hours-based benefits like overtime or the minimum wage. Harris and Krueger argue this new classification would benefit both workers and businesses, reducing expensive litigation by clarifying the rights and obligations of each party.
Alternative work arrangements may on the one hand represent flexibility of the U.S. labor market. If more parents who are primary caregivers can provide labor when it is convenient to them, that may allow them to participate in the labor market when they would not be able to work a full-time job. Similarly, the flexibility may allow students or semi-retired individuals to remain partially attached to the labor market. In fact, 79 percent of independent contractors preferred their arrangement over a traditional job.
At the same time, alternative work arrangements or gig-employment may conceal insufficient labor demand. These workers may be perfectly willing to take a full-time job and are only in these arrangements because they cannot find traditional employment. Only 44 percent of on-call workers and 39 percent of temporary help workers preferred their work arrangements to traditional employment. In this case, these workers may act in a similar manner to unemployed workers or those out of the labor force who would like a job. A large number of such workers might keep downward pressure on wages despite low levels of unemployment, suggesting policymakers need to consider this reserve of semi-employed when assessing the extent of slack in the labor market.
The different preferences for these alternative arrangements are not surprising considering the starkly different wage patterns across these arrangements. Amongst full-time workers, the median usual weekly earnings of those who are independent contractors are just 4 percent less than workers with traditional arrangements. In contrast, earnings are 10 percent less for on-call workers and 41 percent less for temporary help workers.
In the face of declining labor force participation, the number of workers in alternative work arrangements may be a sign of the American labor market providing a variety of opportunities that workers demand. At the same time, these new arrangements likely require different labor market institutions to protect workers as well as new data to properly understand the state of the labor market.
Thank you to Audrey Breitwieser, Ben Delsman, and Yared Lingo for excellent research assistance. Thanks also to Adam Looney and Kriston McIntosh for insightful feedback.
1. Contract workers must work for the same customer and usually at the customer’s worksite. To the extent that many contract workers work off-site, this survey may not accurately capture shifts in this type of employment.
2. This BLS definition includes “wage and salary workers who, based on other than personal reasons such as retirement or returning to school, did not expect their jobs to last, and self-employed and independent contractors employed as such for a year or less who expected to remain self-employed or working as an independent contractor for at most an additional year.”