The Triumph of the Gig Economy Is Postponed Yet Again

When economic times are good, people prefer regular jobs to irregular ones.

Benefits would be better.

Photographer: Andrew Harrer/Bloomberg

Here are two things about the U.S. labor market that are pretty much always true:

  1. It’s changing.
  2. It’s not changing as fast as people say it is. 1

These semi-eternal truths are good to keep in mind while digesting Thursday’s news that the percentage of American workers in “alternative work arrangements” (independent contractors, temp-agency employees, etc.) in 2017 was actually lower than the last time the Bureau of Labor Statistics asked, in 2005.

The U.S. Job Market Hasn't Gone Alternative

Percentage of U.S. workers in alternative work arrangements

Source: U.S. Bureau of Labor Statistics

The percentage of self-employed workers, whom the BLS keeps track of in its monthly employment surveys, is down since 2005, too.

U.S. Workers Aren't Going It Alone, Either

Self-employed workers as a percentage of total employment, not seasonally adjusted

Source: U.S. Bureau of Labor Statistics

These trends in the data surely must seem strange to any regular consumer of business and economics media, which have been bombarding us with news of the rise of the gig economy, the new era of freelancing and the end of the traditional job for a couple of decades now. Heck, they may seem strange to anybody who has taken an Uber or stayed in an Airbnb, and probably to lots of people still struggling to land a permanent job with full benefits. 2

Here are my four main explanations for the disconnect, which I’ve been honing in way too many columns since writing a piece titled “Where Are All the Self-Employed Workers?” in 2014:

  1. These measures are incomplete, and surely miss a lot of real changes in the workplace and labor market.
  2. There’s a big cyclical component at work that people sometimes mistake for a long-term trend.
  3. There have always been tons of Americans doing work other than full-time, permanent “traditional” jobs.
  4. There are sometimes compositional effects in the data that mask important shifts.

The biggest measurement issue is that BLS surveys tend to focus on people’s primary jobs. There is a question in the monthly Current Population Survey about holding multiple jobs, but that data too shows a decline over time. What may be missing is occasional work, be it driving an Uber or helping a friend with a graphic design project or even babysitting the neighbor’s kids. In its 2017 contingent and alternative work arrangements survey, the BLS for the first time asked four questions about “electronically mediated employment” that were designed to get at some of this, the results of which it will publish at a later date. In the meantime, we can look to the Census Bureau’s statistics on non-employer businesses, which are derived from tax returns (these take a while to become available, which is why the most recent numbers are from 2015) and show a significant rise in the number of sole proprietorships and other solo endeavors, which I’ve divided here by total employment to make comparisons over time more meaningful.

The Growing Ranks of Solo Businesses

Non-employer establishments as percentage of total employment

Sources: U.S. Census Bureau, U.S. Bureau of Labor Statistics

Along these same lines, the more timely “Freelancing in America” surveys commissioned by the Freelancers Union and the online talent marketplace Upwork since 2014 show that diversified workers, defined as “people with multiple sources of income from a mix of traditional employers and freelance work,” have been growing in number while the number of independent contractors has gone down.

It's All About the 'Diversified Workers'

Estimated number of U.S. workers by category

Sources: "Freelancing in America" surveys conducted by Edelman Intelligence for Upwork and Freelancers Union

That the number of independent contractors has declined in recent years shouldn’t be all that big a surprise, given that the same thing happened in the latter years of the 1990s expansion, according to the BLS data shown in the first chart. As the unemployment rate falls toward historic lows, employers become more willing to offer jobs with benefits and workers less willing to put up with contract or freelance work. Conversely, the sharp rise in independent work that showed up early in the expansion in the “State of Independence” surveys commissioned by MBO Partners, which provides back-office services for what it terms independent workers, turns out to have been a reflection more of the still-terrible state of the job market than of any epochal shift to independent work. MBO Partners predicted in 2011 that by 2020 “70 million people, more than 50 percent of the private workforce, will be independent,” but its own subsequent surveys found that the number of full-time independent workers actually started falling in 2015, and that while the number of independent workers has kept growing modestly if you include part-timers, at an estimated 41 million people as of 2017 there’s no way it’s going to come close to fulfilling that forecast of 70 million by 2020.

That’s still lots and lots of people, and recent efforts to call more attention to the wants and needs of workers in nontraditional jobs are all to the good. But there have always been lots of Americans doing what today we would call freelance or independent work. In the 19th century, it was most Americans, and as recently as the late 1940s (when the BLS started counting these things), more than 18 percent of U.S. workers were classified as unincorporated self-employed, compared with about 10 percent for combined unincorporated and incorporated self-employed now (the incorporated self-employed data only goes back to 2000). Much of that decline can be attributed to the shrinking of independent employment in fields such as agriculture and retail, though, so it is possible that freelance arrangements have become more common over time in white-collar work.

Such compositional changes can be really important. The BLS measure of median job tenure, for example, has been going mostly up since the 1980s, which goes against the widespread belief that jobs have gotten less secure and workers less loyal. But that belief is not at all misplaced among, say, men aged 45 through 54, whose median job tenure declined from 12.8 years in 1983 to 8.4 in 2016. The overall upward trend in tenure has been driven entirely, two sociologists concluded in 2014, by women becoming more likely to stick with their jobs after having babies.

There are compositional effects at work in the alternative work arrangements data, too, although not huge ones. The percentage of workers in contract, on-call or temp jobs in transportation and utilities rose from 9.4 percent in 2005 to 11.9 percent in 2017. In information, which encompasses the media and the software industries, it rose to 11.4 percent from 9.3 percent. So there is some residue of the Ubers, Lyfts and Upworks of the world in the data after all! But all 12 of the other super-sectors tracked by the BLS showed declines. The gig economy’s reach is quite limited.

This doesn’t mean jobs are what they used to be, or that bigger changes aren’t in the offing. It’s just that, at a time like the present when regular jobs are readily available, people seem to be less enthusiastic about doing irregular ones.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

  1. Self-plagiarism alert! I'm pretty sure I wrote something similar to this at the beginning of another column a few years ago. But (1) I write lots and lots of columns and (2) it's still true.

  2. They may also seem strange to anyone familiar with the smaller-scale replica of the BLS “contingent and alternative work arrangements” survey that economists Lawrence Katz and Alan Krueger commissioned in 2015, which found a significant rise in alternative work since 2005. But I don’t have a good explanation for that particular disconnect, and I imagine Katz and Krueger will attempt to address it in the coming weeks/months.

To contact the author of this story:
Justin Fox at

To contact the editor responsible for this story:
Brooke Sample at

Before it's here, it's on the Bloomberg Terminal.