Why a No-Capital Service Business Is the New Job

In August, Americans filed 531,728 business applications on a seasonally adjusted basis, the fifth-highest monthly total since the Census Bureau began the series in 2004. Of those filings, 34,263 reported a planned first date for paying wages. That second figure was smaller than the equivalent number in August 2019, a month when total applications ran about 82% lower. A near-record filing wave is producing fewer payroll plans than a much smaller wave did seven years ago. That comparison changes how the entrepreneurship boom should be read.

A Filing Is Not a Firm

The Census Bureau's Business Formation Statistics separate four stages that coverage routinely collapses into one: an EIN application, a high-propensity application, an application that plans to pay wages, and an employer business that actually files payroll tax. Only the last of those adds a job to the economy. The pipeline narrows hard at every step, and the narrowing is where the story sits.

July 2026 set the record for applications at 578,926, up 8.1% on June. Of those, 151,857 cleared the bureau's high-propensity screen, a set of characteristics that historically predict a payroll. Just 35,024 planned wages. The bureau's projection for employer formations from the August cohort, meaning businesses that will register payroll liabilities within four quarters, stands at 28,501. That's 5.4% of the applications filed in the month.

Over the twelve months to July, Americans filed 6,141,390 business applications, 15.9% more than in the prior twelve months. The series measures intent with reasonable fidelity. It doesn't measure businesses.

The Wage-Planning Share Has Roughly Halved

Push the arithmetic one step further. If August applications were 82% above their August 2019 level, the 2019 base was roughly 292,000. Wage-planning applications that month exceeded 34,263, which puts the 2019 planning share above 11.7%. The August 2026 share is 6.4%. Roughly one filing in sixteen now anticipates a payroll, against better than one in nine before the pandemic.

Two costs moved at once, and they compound. The cost of the alternative rose, meaning the cost of borrowing to hire, buy equipment, or lease space. The cost of not choosing rose much less, because a service business that needs a phone, a vehicle already parked in the driveway, and a skill can open on a Saturday morning.

Why Capital Is the Expensive Input

The NFIB's September survey of small employers puts the average rate paid on short-maturity loans at 8.2%, up 0.7 points from August and back near the series' long-run average of 8.3%. A net 8% of owners reported paying a higher rate on their most recent loan than on the one before, the highest reading since June 2025. The optimism index slipped to 97.9, close to its 52-year average, with expected credit conditions at net -3% and the share calling it a good time to expand down five points to 7%.

Now the inversion. Borrowing $1,000 at 8.2% costs $82 a year in interest. That's one mid-sized lawn care job, one commercial cleaning visit, one afternoon of bookkeeping. At the entry price point of a service business, the cost of capital isn't what stops anyone. The cost of the first customer is, and no lender prices that.

The First Customer Is the Real Capital

Consider what landing that first customer takes with no money: a week assembling a one-page site, several evenings of cold emailing or door knocking, a set of quotes, a follow-up sequence. Forty hours before the first invoice is a fair estimate for a cold start. Valued at $25 an hour of foregone wages, that's $1,000 of labor spent in order to avoid borrowing $1,000.

I'd argue that's the model in a single line. The no-capital business is financed with unpaid hours, and the shadow price of those hours is the founder's alternative wage. When hiring is soft, that alternative is worth less, so the trade looks better. This is why the wave grew during a stretch of expensive credit rather than in spite of it.

The Stock Confirms the Flow

The Small Business Administration's 2026 profile counts 36.2 million small businesses in the United States, of which 29.8 million, or 82.3%, carry no employees. The remaining 6.4 million run a payroll and hold 62.3 million workers between them, an average of about 9.7 each. Census nonemployer statistics for 2023 put that no-employee count at 30.4 million on a slightly different basis.

Those are stock figures and stocks move slowly, which is why the flow carries more signal. If 93.6% of current applications include no wage plan, the nonemployer share of the business population can only climb. The country is adding business owners faster than it's adding businesses that pay anyone, and the gap has widened every year the application series has run hot.

Britain's Register Shows the Same Churn, Faster

The UK gives a cleaner view of what cheap entry does to a register. Companies House recorded 815,277 incorporations in the financial year to March 2026, up 1.67% on the year before, with 192,287 in the June quarter alone.

The register held 5,479,045 companies at the end of March and 5,516,377 at the end of June, a net gain of 37,332 across a quarter in which 192,287 were created. The residual implies roughly 155,000 removals in three months. For every five companies added, four came off. That churn is the signature of capital-light entry, where the cost of starting sits close to the cost of stopping, so both happen at volume rather than as events.

What It Does to Service Prices

A solo operator with no fixed capital is a price taker. Capacity can enter a local market for the price of a mop and a phone, so the clearing price settles at the marginal provider's reservation wage rather than at a cost-plus margin. That's visible in how solo service income distributes, with a dense cluster of low outcomes and a thin tail of high ones, a shape that shows up plainly in freelance rate data.

For markets, the implication runs two ways. It's disinflationary at the bottom of the local services basket, which is part of why headline services inflation and the price of a house clean have drifted apart. And it's a quiet ceiling on wage bargaining, because a marginal worker with a viable zero-capital alternative doesn't negotiate like someone with none.

The directional view follows from the loan rate. As long as the cost of borrowed money exceeds the expected return on a first hire, the capital-light route wins on expected value, and the application flow stays high while the wage-planning flow stays low. NFIB respondents reporting negative earnings trends, a net -20% in September, are telling you the return isn't there yet. There's a related pattern in how small business owners have been choosing growth over profit, and it points the same direction.

The Variable That Decides It

One number settles whether this is capital formation or labor absorption: the conversion rate from application to payroll, currently 5.4%. If it turns up, the filings are a genuine leading indicator of hiring and the boom deserves its headlines. If it keeps sliding while applications hold near record levels, the country is running a very large, very cheap program for converting unemployed labor into self-employed labor, and booking the output as enterprise.

The Census Bureau publishes that conversion as a projection each month, and it has been drifting down, 4.6% lower for the August cohort than for July's. Applications get the coverage because the number is large and arrives monthly. The planned-wage line is the one that funds a payroll, and in August it was smaller than it was in 2019.

Disclaimer: This article is for informational purposes only and does not constitute financial advice. Past performance is not indicative of future results. Consult a qualified financial advisor before making investment decisions.