Tomorrow’s release of the 2017 Contingent Worker Supplement (CWS) will provide the first government estimate of contingent and alternative work in 13 years. The last time the Bureau of Labor Statistics (BLS) collected this supplement was 2005—the Great Recession hadn’t started, smartphones and apps didn’t exist, and YouTube had just launched. Given how much the economy and technology have changed, the new CWS numbers are long overdue. As there has been much discussion and debate about the changing nature of work, it is important to keep in mind exactly what these new data do and do not tell us about the workforce. Here are nine things to keep in mind about the new data.
1. CWS is only providing information about people’s main jobs.
Since the CWS only asks about people’s main jobs—the jobs where they work the most hours—it leaves out any work that is done to supplement other income sources. Since it only asks about work in the past week, it leaves out occasional work. Other surveys, like the Fed’s annual Survey of Household Economics and Decisionmaking (SHED), ask specifically about work people do on a supplemental or occasional basis. These surveys find that about half of workers who engage in alternative or independent work do so to supplement another income source.[1]
2. The CWS is measuring two different types of work arrangements.
The CWS estimates the number of American workers in contingent and alternative arrangements as their main job. These are two separate categories. Contingent arrangements are short-term or temporary jobs; people surveyed expect them to end within a year.[2] Alternative arrangements are particular types of relationships between workers and the individual or company who pays them: independent contractors, temporary help agency workers, on-call workers, and workers provided by contract firms. What brings these categories together as “alternative” is that workers in all of them are not employed directly or regularly by the company at which they report to work on a daily basis.
Alternative arrangements may or may not be contingent. For example, someone who is hired for seasonal help in retail, paid a wage directly by the store where they work, and expects the job to end when the season ends would be in a contingent, but not alternative arrangement. A freelance graphic designer who had worked for the past two years in this capacity and planned to continue indefinitely would be in an alternative, but not contingent arrangement.
3. The CWS is one of the most rigorous studies of alternative work arrangements.
Although there have been many surveys of alternative and independent work arrangements (see our Gig Economy Data Hub if you want to better understand research in this area), the CWS is one of the most rigorous. It is one of the only studies that uses a nationally representative sample and is not administered online. It also has a higher response rate than other studies because it carries the reputation of the government and the BLS. Consequently, it is more likely than most other surveys to be representative of the entire population—especially populations who are often overlooked, like lower-income workers and those without access to the Internet.
In addition, the questions used by the CWS provide clear explanations of the arrangements they ask about. Where some surveys simply ask, “is your job temporary,” for example, the CWS provides a specific time frame and set of circumstances, and little is left to respondents’ interpretation.
4. But the CWS is still one study in a complex landscape of data.
Although the CWS has many strengths, it should not replace what we have learned about these types of work arrangements from other surveys and analyses. No one data source can paint the full picture of what work looks like. Different studies ask different questions and provide us with different pieces of the puzzle. For example, we can gain additional insights from administrative data, such as tax records and financial activities. Qualitative research can provide insights into the meaning of work in people’s lives. We need a range of data sources to fully address the different ways in which Americans earn their livelihoods.
5. Compare the new CWS numbers against the 2005 release.
When looking at the complexity of data on work arrangements, it is essential to make apples-to-apples comparisons. Many of the reports with larger estimates have broader definitions of alternative or independent work. Most research using private administrative data is limited to one or a handful of specific companies. The best comparison for the new CWS numbers are the last BLS CWS numbers from 2005. Even economists Lawrence Katz and Alan Krueger’s replication of the CWS in 2015 is not a direct comparison, as it used a different sampling method and had differences in question wording. Different studies tell us different things; they need to be put in context with one another without being conflated.
6. The CWS data provide little information on online-platform work.
How many workers use online platforms like Uber, TaskRabbit, or Wag!? The CWS actually offers little insight here. We know from other data sources that many people who work on platforms do so on a supplemental or occasional basis,[3] and that there is high turnover among these workers.[4] Since the CWS only asks about workers’ main jobs in the past week, a lot of this work will not be reflected in the data released this week.
Some online platform workers are likely to be included in the alternative and contingent numbers—so long as their work matched the criteria set by the supplement. Since Uber drivers are independent contractors, for example, their work would be considered an alternative arrangement if they relied on it as their main job in the past week. However, it would only be contingent if they had been doing it for less than a year and expect it to last for an additional year or less. There is no way to identify online-platform work, though, from other types of independent contracting in the released CWS data.
The BLS did ask four questions specifically about online intermediaries on the 2017 CWS, but will not be releasing the data this week and have not yet announced a future release date.
7. The relationship between CWS data and the overall economy is ambiguous.
Would an increase or decrease in the number of contingent and alternative arrangements signal economic growth since 2005? There is not a clear answer. Some labor statistics—like unemployment rates or wage data—have relatively clear relationships to the strength of the overall economy. If unemployment is low and wages are high, the economy is perceived to be healthy. Contingent and alternative arrangements are not as clear indicators. On the one hand, we may expect people to take on stable, full-time work if it is available, meaning alternative arrangements would decrease in times of strong economic growth. However, we may also see people pursue independent work, especially independent contracting, at higher rates when there is an atmosphere of confidence, since the jump into self-employment may seem less risky.
8. We should look at if and how these workers access benefits and protections.
The total number of contingent and alternative workers is not the only one that matters. We also need to look at the experience of these types of work. The CWS includes questions about workers’ access to health insurance and pension plans. Workers, regardless of their work arrangement, should have access to basic benefits and protections. In 2005, both contingent and alternative workers were less likely to access these benefits than other workers.[5] The updated numbers will highlight the needs of these workers for policymakers. Although we might not be able to draw conclusions about the economy as a whole from the CWS, we can identify where there are needs to improve economic security among workers.
9. The 2017 CWS is a starting point.
The new CWS is an essential and overdue addition to knowledge on work arrangements in the U.S. After 13 years, it will provide us an essential reference point. However, it is only the beginning of a better understanding. As the world changes, we need ongoing data in order to track trends and make forecasts. When the CWS was first introduced in 1995, it was administered every two years for almost a decade before Congress defunded it. The 2017 CWS needs to mark a new beginning of regularly and consistently collected public data on work arrangements.
[1] See Upwork and Freelancers Union’s Freelancing in America and McKinsey Global Institute’s 2016 Independent Work report.
[2] The BLS produces three different estimates of contingency: the first includes only wage and salary workers, the second includes wage/salary workers and independent contractors who have been in their job for less than a year and expect to work an additional year, and the third includes wage/salary workers who do not expect their job to last regardless of their current tenure and independent contractors from the second estimate. More information here.
[3] See Thumbtack’s June 2018 A Positive View of the Future of Work report and McKinsey Global Institute’s 2016 Independent Work report.
[4] See JPMorgan Chase Institute’s 2016 The Online Platform Economy report.
[5] Other recent research has shown a similar pattern, including the SHED and Treasury Department analysis of tax records.
