There's a pretty good chance that you, or someone you know, has at some point worked a job that the US Bureau of Labor Statistics (BLS) would classify as "contingent" or “alternative.” These are jobs that, respectively, are not expected to last and that do not come with a traditional employment relationship. In its most modern form, this can include gig work, such as driving for Uber or providing freelance labor through TaskRabbit. However, the majority of contingent and alternative work is comprised of independent contracting, work for contract companies, on-call work, and work for temporary help agencies. On Thursday, June 7, BLS will release the results of its first survey of contingent and alternative work since 2005, giving insight into how the incidence of these types of work has changed over the past 12 years. To add context to the highly anticipated data release, let’s look at the results of the 2005 survey and outline what the release on Thursday will (and won’t) tell us.
In February 2005, questions about contingent and alternative employment were added to the regular monthly Current Population Survey (CPS). Specifically, in the case of contingent work, CPS respondents were asked whether they expect their job to continue, and for how long, or whether it is temporary in nature. BLS reports that, in February 2005, contingent workers comprise between 1.8 and 4.1 percent of the workforce. These estimates vary based on the definition of contingent — the low estimate of 1.8 percent excludes the self-employed and independent contractors, and persons who expect to continue their job for more than a year. Long-term independent contractors and self-employed workers (those who have held the job for more than a year or expect to continue in the job for more than a year) are also excluded from the larger estimate of 4.1 percent of the workforce.
As reported by BLS, contingent workers in 2005 were more likely than the overall workforce to be under age 25 and, relative to noncontingent workers of the same age, were more likely to be in school. Contingent workers were also less likely to be white, compared to their noncontingent counterparts. Contingent workers were less likely than noncontingent workers to have employer-provided health insurance or to be eligible for employer-provided pension plans. More than half of workers in this category said that they would prefer noncontingent work, and about a third said that they preferred their current arrangement.
While the nature of some work is itself contingent, a potentially more impactful component of the 2005 survey identifies people who are in alternative work arrangements, including on a long-term basis. These workers can fall outside of traditional social insurance protections and often receive lower pay and fewer benefits (particularly health care and retirement benefits) than similar workers with traditional arrangements. For these reasons, employers have profit motive incentives to increase the incidence of this type of work, particularly when workers have diminished bargaining power. BLS reports that in February 2005, 7.4 percent of workers were independent contractors (up from 6.4 percent in 2001), 1.8 percent were on-call, 0.9 percent were temporary help agency workers assigned to other workplaces, and 0.6 percent were contract company workers assigned to other workplaces.
There are important differences between these four alternative work categories. In February 2005, independent contractors tended to be older (median age of 44) and were more likely to be white (80 percent of the total) and male (64.7 percent of the total). On-call workers had a median age of 39, and about half were women. Temp agency workers had a median age of 33, and the majority were women (56.2 percent). Finally, contract company workers had a median age of 36, and were the most likely to be men (69 percent).
The regular monthly CPS can already give some insight into trends since 2005 for two of the alternative work categories. Regularly published data show that self-employed workers (both incorporated and unincorporated ), which can serve as a proxy for independent contractors, make up a slightly smaller share of the overall workforce than they did in 2005. Similarly, employment in temporary help services (which includes workers who are not assigned to other workplaces) is practically unchanged since 2005 in its share of the total workforce. What we do not know, is whether there are upward trends for contract company workers assigned to another workplace and on-call work.
In particular, if the incidence of work for contract companies that provide workers to another workplace (a subset of subcontracting) has increased, it has ramifications for health and retirement security and could play a role in explaining broader changes to the labor market. In the 2005 survey, subcontracting was most common among IT workers, security guards, technicians, and cleaning and building service occupations. One question that may be answered on Thursday is whether other industries and occupations have started to follow the subcontracting model. Separately, the release can also offer insight into whether the incidence of subcontracting has increased within the industries and occupations that already used this approach in 2005.
Importantly, the perhaps most widely anticipated component of the May 2017 survey, the measure of gig work, will not be released on Thursday. In announcing the release of the first set of survey results, BLS reports:
“Four new questions were added to the May 2017 Contingent Worker Supplement. These questions were designed to identify individuals who found short tasks or jobs through a mobile app or website and were paid through the same app or website. These data will not be included in the June 7th release. BLS continues to evaluate these data and plans to publish the findings from this research at a later date.”
The overall size of the gig workforce is debated in economic literature, with some indications that it is, at any given point in time, a fairly small portion of the total workforce. A 2015 survey conducted by Katz and Krueger shows that subcontracting has increased the most since 2005 (from 1.4 percent to 3.1 percent of the workforce), but that gig work makes up only 0.5 percent of workers. Separately, there does seem to be some indication that gig work, often done as a supplement to other streams of income, is more common when measured over a one-year time frame, as would be shown by tax returns, than over the one-week time frame captured by the CPS. In other words, part-year gig work may be much more common than full-year gig work. Those interested in BLS analysis of gig work will have to wait a bit longer.
Readers may wonder why this survey was not conducted from 2006 to 2016. Budget cuts to US statistical agencies, including BLS, forced the agencies to make tough decisions. From 1995 to 1999, the contingent work supplement to the CPS was conducted every other February. The results of these three surveys were remarkably similar. Analysis showed very little trend in the incidence of either contingent or alternative work. With statistical agencies facing inflation-adjusted budget cuts, the supplementary questions were asked to only three quarters of CPS respondents in 2001, the 2003 survey supplement was not conducted, and again in 2005 only three quarters of the CPS sample received the supplementary questions. While there were indications that some categories of alternative work were increasing by 2005, BLS continued to face a shrinking real budget. Since 2005, the BLS budget has fallen 10 percent in inflation-adjusted terms.
There is a lot to look forward to in this week’s release from BLS. While the new data won’t immediately answer all of the questions people have about contingent and alternative work arrangements, it will offer the first official insight in a dozen years. As outlined here, this insight is particularly important in the case of subcontracting. At a later date, presumably later this year, BLS will release additional information on gig work, and the Census Bureau will release the underlying data that was used by BLS to produce its analysis. Those interested in digging deeper into the subject should stay tuned for CEPR’s detailed analysis of the underlying data.
 The category of workers who are self employed and unincorporated includes gig workers.