Contingent Work In The U.S. Labor Market

For the past 20 or so years, it is fairly common to hear that the U.S. workforce is an increasingly precarious workforce – that is, made up of jobs with non-standard employment arrangements, such as temp jobs, on call work, and independent contracting.

Because these types of employment arrangements, often called “contingent work,” tend to offer less stability, lower wages, and less opportunities for advancement, compared with “standard” full-time jobs, the growth of the contingent workforce is often portrayed as a cause and/or signal of the erosion of workers’ rights and the decline of the middle class in the U.S. Others see it differently, however, and argue that contingent work offers the flexibility desired by employers and employees alike, and that flexible jobs allow faster and more efficient “matching” of workers with positions, thus boosting productivity. This debate, of course, centers largely around empirical questions, and the body of research on contingent work has been building for a few decades now (see Kalleberg 2000Connelly and Gallagher 2004). Yet not all labor force surveys are designed to capture the full set of nuances of workers’ employment arrangements. Starting in the mid 1990s, the Bureau of Labor Statistics (BLS) had the good sense to collect data on this topic, in the form of the Contingent Worker Supplement (CWS) to the Current Population Survey (CPS). The CWS was administered five times between 1995 and 2005, and provided valuable data on these “nonstandard” employment relations.

The CWS, however, has not been conducted since 2005, substantially decreasing the high quality information available on contingent work at a particularly important time, given that the Great Recession began shortly thereafter. The U.S. Government Accountability Office (GAO) made a laudable attempt to fill this hole with an April 2015 report, which uses several data sources to provide an important snapshot on the prevalence of and trends in contingent work in the U.S. (the data go up to 2010). There are a few key takeaways from this report, some of which are long established.

The size of the contingent workforce, as well as its implications, vary dramatically depending on how one defines it. The broadest possible definition of contingent work includes the following types of arrangements:

  1. Temporary workers employed by temp agencies;
  2. Temporary workers employed directly;
  3. On call workers and day laborers;
  4. Contract company workers;
  5. Independent contractors;
  6. Self-employed workers;
  7. Part-time workers.

Types 1-4 are often called “core contingency” workers, as these are the jobs that most clearly embody unstable, “non-traditional” employment arrangements (Polivka 1996). According to the GAO, “core contingent” workers constitute approximately 7-8 percent of the 2010 U.S. workforce. These workers tend to be younger and less educated, and they are disportionately Hispanic.

Using a broader definition of contingent work, of course, increases the size of this group, which we might characterize as workers with some kind of alternative employment arrangement. In 2010, around 16 percent of workers were either independent contractors (12.9 percent) or self-employed (3.3 percent), while another 16.2 percent reported working part time. So, using this most inclusive definition, about two in five U.S. workers (40.4 percent) had some sort of employment arrangement other than standard full time employment.

Without question, the definition of contingent or alternative work is a difficult issue, particularly when it is used in political discussions. For example, part time workers tend to have little employment security and scant if any benefits, yet the majority say they are working part time by choice, thus complicating somewhat the argument that these workers represent a decline in U.S. workers’ rights and conditions. In contrast, according to the GAO, most temporary workers would prefer regular full time employment but cannot find it, which seems to support the argument that standard, stable employment is being denied many workers. Yet even a substantial proportion of temps (about one third) report that they prefer their current arrangement.

Clearly, then, people are in these jobs for different reasons, and generalizations should be made with caution, but a meaningful proportion of U.S. workers occupy these non-traditional positions.

Contingent work is associated with lower pay and benefits, and less job stability. According to the GAO estimates, even controlling for factors such as demographics, education, union membership, industry/occupation, and geography, contingent workers’ hourly wages are about 10 percent less than those of standard full time workers. They are also substantially more likely to experience job separation (e.g., unemployment), to report having no employer-provided benefits, and to experience poverty. These findings square with prior research (e.g., Kalleberg et al. 2000)

Some of these outcomes, interestingly, vary by occupation and industry. For example, contingent workers in the construction industry earned wages similar to those of their regular full time counterparts, whereas they earned less in other industries, including education and transportation.

The size of the core contingent workforce increased modestly during the Great Recession. Since, as mentioned above, the GAO uses data sources alternative to the CWS, comparisons over the long term are difficult, but between 2006 and 2010, the size of the “core contingent workforce” appears to have increased, from 7.1 to 7.9 percent, including a 0.4 percentage point increase in temp agency workers’ share (0.9 to 1.3 percent), and a one percentage point increase in the share of on call workers (2.5 to 3.5 percent). Due to a 0.6 percentage point decrease in the share of contract workers, the total size of the core contingent workforce increased about one percentage point during these peak years of the Great Recession.

Perhaps the most striking change, however, can be found in the proportion of part time workers, which increased from 11.9 percent in 2006 to 16.2 percent in 2010. This jump is almost certainly due in part to the Recession.

Needless to say, these trends are affected by unemployment (and workers who leave the workforce entirely). For example, the size of the contingent workforce might have grown more if a large proportion of workers who would have had contingent arrangements were unemployed or left the labor force.


Broadly defined, contingent workers are a diverse group. They range from self-employed business owners to part timers and temps. And, even among “core contingency workers,” the situation is not straightforward, as many of them work in these jobs voluntarily, for the flexibility and other reasons.

That said, many contingent positions offer less compensation and little stability to workers who would prefer regular employment. This is a fundamentally different employee/employer relationship from regular full time employment, and the proportion of U.S. workers who are in similar positions is almost certainly higher than the numbers above suggest. That is, there are most likely millions of workers whose positions are unstable but who are not classified as contingent workers. Moreover the contingent workforce may affect the situation of regular, full time workers (e.g., Pedulla 2014)

Now, again, there are different ways to view this segment of the workforce. Some see it as exploitation whereas other focus on the benefits of flexibility and better job matching. This is an important debate, yet the BLS is no longer collecting the detailed data that would be very helpful in addressing these questions. The CWS has not been authorized for over a decade. Regardless of how you view contingent work, there is no question that the workers in these positions represent a significant share of the U.S. workforce, one that is unique in many respects, and one that can only be monitored and understood if we collect more information necessary to do so.

Issues Areas

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and email addresses turn into links automatically.
  • Lines and paragraphs break automatically.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
2 + 2 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.