You can see the gig economy everywhere but in the statistics.
For years, economists, pundits and policymakers have grappled with the rise of Uber, the growth of temporary work and the fissuring of the relationship between companies and their workers. Optimists cheered the flexibility offered by the freelance life. Pessimists fretted about the disappearance of traditional jobs, with the benefits and legal protections they provided.
That debate has played out largely in the absence of solid data. But on Thursday, the Bureau of Labor Statistics released its first in-depth look at nontraditional work since 2005, and came to a startling conclusion: The old-fashioned job remains king.
Roughly 10 percent of American workers in 2017 were employed in some form of what the government calls “alternative work arrangements,” a broad category including Uber drivers, freelance writers and people employed through temporary-help agencies — essentially anyone whose main source of work comes outside a traditional employment relationship. Far from a boom in gig work, that represents a slight decline from 2005, when about 11 percent of workers fell into those categories.
“I think everybody’s narrative got blown up,” said Michael R. Strain, director of economic policy studies at the American Enterprise Institute, a conservative think tank.
Mr. Strain and other experts cautioned that the data released on Thursday did not signal the American workplace had remained static over the past decade. The government’s numbers, by design, do not include people who do gig or freelance work in addition to traditional jobs, and they may not fully capture income-generating activities that people might not consider “work,” like renting out a home on Airbnb.
Separate data released by the Federal Reserve this month found that nearly a third of adults engaged in some form of gig work, either as a primary job or to supplement other sources of income. Private-sector surveys have reached similar conclusions.
Nor does the bureau’s data reflect other changes that have left many American workers with less security and fewer opportunities for advancement. Many companies, for example, now outsource large parts of their business to subcontractors. Employees of those firms will not, for the most part, count as alternative workers under the government’s definition. But they generally earn less and receive smaller benefits than equivalent workers employed directly by large companies, and they have far less opportunity to move up the corporate ladder.
“In my view, it’s this domestic outsourcing that is the big change in why wages don’t rise and why workers feel so insecure,” said Eileen Appelbaum, co-director of the Center for Economic and Policy Research, a liberal think tank. “And we have only the most indirect data to show it.”
Rafael Sanchez is one of those who may fall into a gray area in the statistics. He moved to the United States from Mexico 16 years ago, eventually settling in New Brunswick, N.J., and was initially able to find steady work, including a five-year stint as a full-time employee of a window factory.
But Mr. Sanchez was laid off in 2007, and he has since worked for a series of temporary-help and staffing agencies, moving from factory to factory and warehouse to warehouse, earning $ 9 an hour to pack boxes and do other manual labor. The positions can last months or even years without turning into permanent jobs working directly for the factories, which offer better pay and room for advancement.
“They’re not jobs where you can get ahead,” Mr. Sanchez said through an interpreter.
Economists have long argued that the most visible kinds of gig work are a relatively small part of the overall labor market, and that nonstandard work arrangements long predated the emergence of app-based platforms like Uber and TaskRabbit. (Uber’s impact was somewhat visible in the data: The number of independent contractors in the transportation and utilities industry increased by about 200,000 from 2005 to 2017.)
The report on Thursday did not break out online workers; a follow-up report scheduled for September will do so. But it portrayed an alternative work force far broader than the one typically conjured by the term “gig economy.”
The largest category of alternative workers, independent contractors, are disproportionately in their mid-40s or older and common in sectors like construction that have not been disrupted by Silicon Valley entrepreneurs. They earn about as much, on average, as standard employees, and are relatively happy with their arrangements: Nearly eight in 10 say they prefer being an independent contractor to being an employee.
The picture looks different for on-call and temporary workers, who tend to be younger, earn less and rarely have access to employer-provided health or retirement benefits. They are also far more likely to say they would prefer traditional work.
Neither category, however, has grown as a share of the work force since 2005. Tom Gimbel, chief executive of LaSalle Network, a staffing firm in Chicago, said the narrative of the gig economy had long been overblown. “Because of social media and stories out there, I think the perception of the contingent work force has been inflated,” he said.
Still, the data left many economists perplexed. The Bureau of Labor Statistics, for budgetary reasons, had not conducted its survey of alternative work since 2005. But in 2015, the economists Lawrence F. Katz and Alan B. Krueger conducted their own survey that tried to mirror the bureau’s methodology. They found that alternative work was more common — and growing far more quickly — than the data released on Thursday suggests.
Economists offered various explanations for the discrepancy. One possibility is that the boom in gig-type jobs was real but that as the economy has improved, more people have been able to find traditional work. Part-time work, which surged in the recession, has fallen in the recovery, and employment by temporary-help services has leveled off. If true of alternative work more broadly, that would suggest that what many commenters interpreted as a structural shift in the economy was instead a temporary result of a weak labor market.
It is also possible that the new data understates real changes in the nature of work. The government’s standard tools for measuring employment have struggled to capture the shifting employment landscape. For example, the Current Population Survey, the monthly survey used to calculate the unemployment rate and other key measures, shows that self-employment is falling, even as tax data from the Internal Revenue Service has shown the opposite.
“The questions on our standard surveys don’t probe into the nature of these arrangements,” said Katharine G. Abraham, a University of Maryland economist who served as commissioner of the Bureau of Labor Statistics under President Bill Clinton. “We’re not asking the right questions, and they’re hard to answer anyway.”
The bureau’s data comes from an add-on to the Current Population Survey intended to fill in some of those gaps. But it has shortcomings. The extra questions, for example, were asked only of people considered “employed” in the standard survey. That might leave out people who earn income through activities that they do not see as a job, such as selling products online or working erratically as a freelancer.
Professor Abraham said there was evidence that workers struggled to accurately report and classify work that did not fall neatly into traditional buckets. Some Uber drivers, for example, might consider themselves employees of the ride-sharing company, even though legally classified as independent contractors. Some Amazon warehouse workers might report being employees of the e-commerce giant, even if employed by an outside staffing firm.
Professor Abraham and other experts said, however, that if there had been a big rise in alternative work, it should have shown up in the latest survey. The fact that it did not, they said, suggested that any shift had been relatively modest.
“Definitely, the vast majority of individuals are still W-2 employees working in traditional jobs,” Professor Katz said. He said it was too early to account for the divergence between the government findings and his study with Professor Krueger.
Teasing out exactly what has and has not changed will require more data, which may not be coming. The Bureau of Labor Statistics in 2016 received funding to revive its survey on alternative work, but the Trump administration has not requested funding to run the survey on a regular basis.
“The fact that our last data point on this was in 2005 makes it so hard to figure out what’s going on,” said Martha Gimbel, director of economic research for the job-search site Indeed. “Measurement is important, and this is why it’s important to fund data analysis.”