About UsArrow Down
This is some text inside of a div block.

Report: Why estimates of the freelance economy disagree

Adam Ozimek, chief economist at Upwork

  • Estimates of the size of the freelance economy come to significantly different conclusions, ranging from some estimating one in ten workers participate to one in three.
  • Using both Bureau of Labor Statistics (BLS) and Freelancing in America (FIA) data, we show that one major reason why there is such a discrepancy in estimates is due to how occasional freelancers are counted (or not).
  • This research compares estimates, examining complexities in counting caused by varying work frequency among different types of workers (such as the self-employed and multiple job holders).

If one were to Google the size of the freelance economy, they would likely get wildly different results. Estimates of the freelance economy can range anywhere from one in ten workers participating to one in three. So what gives? Why do we see such a disparity in estimates? And which estimate is the most accurate?Although working full time for a single employer in a traditional job remains the dominant form of work today, the umbrella of freelancing covers a highly varied array of work types. You can work for one employer who contracts you out or be a temporary worker hired either directly or as a contractor through a staffing firm.

You can be self-employed and work alone, or hold a traditional job and freelance on the side. You can also be part of an agency team of contractors. While the rise of online platforms has made freelancing more productive and accessible, self-employment is not new. It has always been a feature of the modern economy, but one that researchers have struggled and disagreed on how to measure.Start with Freelancing in America (FIA), a survey co-commissioned annually by Upwork and Freelancers Union and conducted by Edelman Intelligence, in 2017, FIA found that 36% of working Americans freelanced. In contrast, the 2017 Contingent Worker Supplement from the Bureau of Labor Statistics (BLS CWS) found that 10.1% of workers were in what they called “alternative work arrangements.” So what is it about the freelance workforce that makes it so difficult to measure? Both FIA and the BLS CWS included multiple types of non-traditional workers, and some of the disagreement between the two is definitional.

For example, the BLS CWS asked only about the main job a person holds, while FIA asked about all work.A major but under appreciated source of disagreement, however, comes from how sources are accounting for those who engage in the freelance economy on an occasional basis.Accounting for “occasional freelancing,” can explain 80% of the gap between the BLS and FIA estimates of the number of independent contractors.

Focusing on Independent Contractors to Understand the Differences

Independent contractors are a subset of freelancers that both FIA and the BLS CWS specifically looked at, which provides an apples-to-apples comparison we can take as an example. FIA estimated 11% of the overall U.S. workforce were independent contractors in 2017, while the BLS data from the same year reported 6.9% for the same workforce segment. In other words, FIA found 1.6 independent contractors for every one found by the BLS, even though both are focused on the same specific type of freelancer.It’s not that one survey is right and the other is wrong. Rather they are asking different questions. FIA asks about any work happening over the previous year, while the BLS asked about work done in only one specific week.

For jobs where workers tend to be employed year-round, this wouldn’t make much difference. However, the choice of asking about only one week instead of asking about activity within one year creates a significant difference for workers who are freelancing given how common “occasional freelancing” is.What is “occasional freelancing?”

Occasional freelancing is when someone opts to freelance once, or at multiple times throughout the year, but not on a daily or weekly basis. The magnitude of occasional freelancing can be seen in FIA data on independent contractors. The survey asks when the most recent freelancing was done, and nearly one-third of freelancers hadn’t freelanced within the last week.

If FIA had instead only counted independent contractors who reported having ongoing work or work in the last week, they would have found 7.8% of the overall U.S. workforce were independent contractors. This number is far closer to the BLS estimate of 6.9% and suggests that occasional freelancing explains 80% of the gap between the two estimates.

Occasional freelancing becomes an even bigger measurement challenge when looking at types of freelancers other than independent contractors. Among temporary workers, for example, nearly half did not participate within the last week.

Given the high share of occasional freelancing among temporary workers, it is easy to see how different surveys can arrive at different answers about the size of this group, depending on the time period they look at.Another type of freelance worker that contrasts vividly with temporary workers is freelance business owners, 80% of whom worked in the last week according to FIA. This illustrates a key challenge in measuring the freelance economy: some freelance work is truly occasional gigs, while other work in many ways looks more like traditional, full-time employment.

Occasional Work in BLS Data

The impact of not taking occasional work into consideration can also be seen in the BLS data.  The Current Population Survey (CPS) produced by the BLS measures two types of non-traditional work arrangements: multiple job-holding and self-employment. While not all freelancers fit into either of these categories, and not everyone in these categories is a freelancer, both have significant overlap with freelancers. As with the BLS CWS, the CPS only counts those who report doing either type of work in a particular week. One advantage of the CPS is that it surveys the same individuals repeatedly over the course of eight separate months in a period of just over a year. This allows occasional workers to be counted by using the percent who are self-employed or hold multiple jobs at any time within those eight months instead of in just one week.

The share of either self-employed or multiple jobholders at any time is significantly higher than the share of workers doing so within the most recent week. Multiple jobholding is 80% higher, and self-employment is 60% higher. Altogether, 26.8% of employed people were either self-employed or a multiple jobholder at some point within an eight month survey period. Compare this to the count using only the shorter one week timeframe, which found that the share participating was far lower at 16.8%. These results show once again how important controlling occasional work is when measuring non-traditional work arrangements.

What This Means for Understanding Freelancing

While the BLS surveys and FIA may have different headline numbers for the size of the freelance economy, beneath the surface, all of these surveys come to important agreements. First, when focusing on the same recent timeframe, the BLS CWS suggests independent contractors make up 6.9% of the workforce, and FIA suggests a close estimate of 7.8%. In other words, both FIA and the BLS CWS estimates are in agreement when looking at the same workers over the same time period. Looking over the entire year, 11% of the workforce does some independent contractor work—a substantially higher number.

In addition, the CPS data illustrates the importance of using a longer timespan to measure how many participate in non-traditional work arrangements. Looking within an eight month period, CPS data shows 26.8% participating in either self-employment or multiple jobholding. In contrast, using the shorter timeframe from the headline CPS only shows 16.8% participating. On a higher level, these data sets illustrate that many freelancers, self-employed, and multiple-jobholders engage in this work on an occasional basis.

As a result, whether people are counted as participating in non-traditional work arrangements will depend on how long of a period of time a survey looks at. How many people freelanced last week is a very different question than how many participated last year.

Given the widespread engagement in occasional freelancing, it is important that analysis of these labor markets incorporate surveys like FIA since they look beyond recent work activity to gain a more complete assessment of activity. With around a third of all workers participating in the freelance economy in some way, this workforce represents a substantial part of the economy that we should be measuring accurately.