Gig Workers Are Twice as Likely to Start Businesses, IRS Tax Data Shows

Gig workers are about twice as likely to start their own businesses as comparable workers who never gigged, and they overwhelmingly start them in the same line of work they gigged in. That's the headline finding from an NBER working paper by Matthew Denes (Carnegie Mellon), Spyridon Lagaras (University of Illinois), and Margarita Tsoutsoura (Washington University in St. Louis), built on administrative U.S. tax return data covering more than 10 million Americans who earned gig income between 2012 and 2021.

For anyone running an appliance repair or HVAC shop, this paper describes something you've probably already watched happen. The tech picking up weekend jobs on Thumbtack or running dispatches through Angi isn't just supplementing income. Statistically, that person is learning the business and getting ready to compete with you.

What the Tax Returns Show

The numbers are specific because the data is. Working from tax filings rather than surveys, the authors found gig workers were roughly 1 percentage point more likely to start a new business, which sounds small until you realize it doubles the baseline rate of entrepreneurial entry. About three-quarters of that effect came from first-time entrepreneurs, people who had never owned a business before gigging.

The effect was strongest among younger workers, lower-income workers, and people who value flexibility, including single parents. And the firms they start aren't random. Gig workers launch businesses in industries that match their platform experience, which the authors read as evidence of on-the-job learning: the gig is functioning as an unpaid apprenticeship in customer acquisition, pricing, and scheduling.

The startups themselves look different too. Gig-founded firms launched roughly 23% larger by revenue and with 39% more employees than comparable new businesses, and their founders tapped debt financing at higher rates. They were also 9 to 18% more likely to use independent contractors themselves, which figures. You hire the way you were hired.

Survival is the catch. Gig-founded firms were about 3 percentage points less likely to make it through their first three years. But the ones that survived grew faster and performed better than firms started by non-gig founders, and founder income ran 13.1% higher three years after launch.

The Platform-to-Shop Pipeline in Home Services

The study covers the whole gig economy, rideshare and delivery included, so the home services slice is our extrapolation, not the paper's. Still, the mechanism maps cleanly onto this trade. A tech running lead-gen platform jobs learns exactly what an employed tech never sees: what customers actually pay, what leads cost, which zip codes convert, and how brutal the platform's take is.

That last lesson is usually the push. Once a tech has done 200 Thumbtack jobs, the math on going fully independent writes itself.

For established shops, the paper cuts two ways. The threat is obvious: a steady stream of new entrants who already know the local market and start hungrier and cheaper than you. The lower three-year survival rate is cold comfort, because even a shop that fails in year two spends those two years undercutting your service call fee.

The opportunity is less obvious. These founders launch bigger, borrow more, and hire contractors early, which means some of them will hit the scaling wall fast. That makes them acquisition targets, subcontracting partners for your overflow, or, frankly, your next experienced hire when the business folds. A shop owner who stays friendly with the new entrants in their market has options the hostile one doesn't.

It also reframes the platforms themselves. Angi and Thumbtack aren't just middlemen skimming lead fees; per this data, they're the largest informal business school in the trades. Anyone weighing the jump should read our guide to starting an appliance repair business in California first, and anyone who already jumped and is drowning in demand should see scaling from solo tech to multi-truck operation.

The paper is slated for the Journal of Financial Economics, so the peer-review box is checked. The pipeline it documents is already running in your market either way.

Source

Denes, M.R., Lagaras, S., and Tsoutsoura, M. (2025). "Entrepreneurship and the Gig Economy: Evidence from U.S. Tax Returns." NBER Working Paper No. 33347. https://www.nber.org/papers/w33347