First, Define the Business: What “Solo” Means in the Data
Before asking why a business does not hire, we need to define which businesses the evidence is counting. This chapter separates nonemployer status from owner income, growth intention, and business survival, then shows how the answer changes when the population moves from broad tax-linked activity to narrower samples that look more like substantive startups. The payoff is simple: a transition rate is meaningful only when its denominator is visible.
Nonemployer does not mean economically empty
In the statistical definition used by the geographic study, a nonemployer business has no paid employees, earns at least $1,000 in annual receipts, and is subject to federal income taxes. It can have one owner or several. A professional contractor, an online retailer, an independent freelancer, or a family business whose relatives help without being on payroll can all fit the definition. What unites them is not their motivation, scale, or legal form. It is the absence of paid employees requiring payroll taxes.
That last condition is narrower than the everyday meaning of “working alone.” A nonemployer may use unpaid family labor. It may hire outside services without putting those providers on its payroll. It may also be one part of an owner’s working life: a side business, a short-term contract stream, or an activity that supplements a wage job. Conversely, a business can be the owner’s primary income and still have no intention of hiring. The category describes an employment arrangement, not a complete account of the business’s value or the owner’s circumstances.
The distinction matters because the question “Why does this business not hire?” contains several possible questions. Is there no paid employee because the owner prefers a flexible solo model? Because demand is too irregular to support a wage? Because the business has not yet reached the intended scale? Because the owner expects to hire but has not found the right worker? Or because the activity has already ended? Payroll status alone cannot tell us which explanation applies.
The three supplied studies observe different slices of this world. The geographic study compares annual counts of nonemployer businesses with employer businesses across U.S. counties from 2000 to 2019. It is useful for showing the size and location of the nonemployer population, but its county-level counts do not follow each individual business through time. A county can have more nonemployers because existing owners stayed active, new owners entered, businesses moved, or some combination of these processes.
The startup-dynamics study uses the Comprehensive Startup Panel, which joins a longitudinal panel of employer establishments with a longitudinal panel of nonemployer establishments. In that study, a startup is identified by the first year in which a nonemployer records revenue or an employer records payroll. The 2011 and 2012 cohorts are then followed for up to seven years after startup, allowing the researchers to distinguish a business that begins without employees and later hires from one that never enters the employer records.
This construction makes the first employee observable as a state change, but it also creates a boundary around the result. The study’s broad universe includes every observed business activity that files taxes. That is valuable for measuring the full population, yet it also brings in small-scale consulting, contracting, and secondary activities that may never have been intended to become employers. A low transition rate in this universe may partly reflect the fact that many counted activities were never candidates for payroll expansion.
The third study begins earlier in the formation process. It links about 17 million applications for Employer Identification Numbers filed between April 2017 and December 2021 to later employment records. Applicants report the highest number of employees they expect within the next twelve months. The study calls employer entry the appearance of positive employment in the linked records. For its job-creation question, a nonemployer business and a business that never materializes are both recorded as zero employment generation.
That is a coherent choice for studying whether an application produces paid jobs. It is not a judgment that a nonemployer produces no economic benefit. A nonemployer can generate the owner’s income and provide flexibility even when it adds no payroll employment. The measure answers “Did this application lead to employer employment within the observation window?” It does not answer “Was the owner’s business worthwhile?”
The same care applies to survival. In the startup-dynamics study, survival means that the business remains active in the relevant Census panel—through revenue in the nonemployer panel or payroll in the employer panel. It does not necessarily mean profitability, full-time work, rising income, or personal satisfaction. A business can survive without hiring, and a business can hire briefly without creating durable employment. These are different states and should not be folded into one success label.
The nested structure in the diagram is conceptual rather than a claim that every category is a strict subset of the next one. The intention groups come from a survey classification reported by the geographic study, while the broad, EIN-based, and substantive groups come from the startup-dynamics analysis. They answer related questions, but they are not interchangeable labels for the same businesses.
The intention behind “no employees” changes the denominator
The geographic study reports a Federal Reserve classification that makes the mixture of intentions explicit. Roughly 30% of nonemployer businesses provide supplemental work and have no future growth intention. About 21% are contract workers for whom the business is the primary source of income but who do not intend to hire. Another 26% are “stable nonemployers”: the business is the owner’s primary income, yet there is no plan to add employees. The remaining 24% are potential employers that intend to hire.
The exact categories should be read as a snapshot of reported intentions, not as permanent types. An owner may change from part-time work to a full-time business, discover that hiring is unnecessary, or postpone a plan after a disappointing season. Still, the classification supplies a missing variable in the payroll data. If nearly three-quarters of the reported categories are supplemental, contract-based, or stable solo arrangements without a hiring intention, then a low employer-transition rate is not surprising. Many of the businesses in the denominator are not trying to cross that boundary.
This is why “most solo businesses never hire” can be true without implying that most owners are failing to execute a growth plan. The statement may include an owner who deliberately sells a specialized service alone, an independent contractor replacing an unavailable wage job, and an aspiring employer who has not yet secured enough predictable demand. All three have zero paid employees, but the relevant constraint and the relevant judgment differ.
The startup cohort shows how a broad denominator produces a very low transition share. Across 8.752 million businesses that began in 2011 or 2012 and were observed for eight years, 89.1% started and ended as nonemployers, while 1.8% started as nonemployers and transitioned to employer status. The remaining businesses began as employers, with some showing positive net employment growth and others not. Those figures are important, but they do not mean that only 1.8% of every kind of solo business will ever hire. They describe this cohort, this definition of startup, and this observation period.
The authors also examine a narrower proxy for businesses that look more like organized, ongoing ventures. Only about 2.3 million of the 8.75 million cohort businesses began with an Employer Identification Number, and roughly two-thirds of those were nonemployers. Of the nearly 160,000 nonemployers classified as having transitioned, only about 25,000 began without an EIN. The overall transition rate was about 2%, but the rate among nonemployers that started with an EIN was 8.9%.
That difference is not a statistical trick. Removing or isolating businesses without an EIN changes the population’s composition. It likely removes many activities that are small, supplemental, or less formal, while retaining businesses whose owners obtained an EIN for banking, licensing, compliance, or other reasons even though they had no employees. An EIN is therefore a useful signal of organizational commitment or administrative need, not a definitive test of whether a business is “real.”
The narrowest analysis in the startup-dynamics study goes further. Its “substantive firms” have some financial capital investment, an owner who works at least 20 hours per week in the business, and a business founded by the current owner. These criteria describe a more committed and resource-bearing subset, but they also select for precisely the traits that may make transition more likely. A higher transition rate in that group should be understood as evidence that the path is different for businesses with more capital and owner time—not as evidence that every owner can obtain the same outcome by meeting an arbitrary checklist.
The denominator also changes because the studies observe different moments. The broad startup panel begins when a business first records revenue or payroll. The application study begins when a potential entrant files for an EIN and asks about expected employment during the next year. An application can lead to a nonemployer business, a delayed employer, an employer that appears within the window, or no observed employer employment. The geographic study counts businesses at the county level rather than classifying each owner’s plan. A percentage from one stage cannot be substituted for a percentage from another.
The application data make the same point from the opposite direction. Among all applications with an intent to start a new business, many report zero expected employment. These applications may reflect a plan to operate as a nonemployer or uncertainty about when hiring will occur. Applications with a planned wage payment have a much higher employer-transition rate—27% within four quarters compared with 7% for all applications in the study—and almost all of them report positive expected employment. The planned-wage group is therefore not a random sample of all applicants; it is a more concrete hiring-intent group.
A reader who sees “7%” and “27%” without the labels might conclude that one study contradicts itself. It does not. The figures compare all applications with a selected group that has already signaled a planned wage payment. The difference is evidence that intent contains information, but it is also evidence that conditioning on intent changes the denominator. The same reasoning applies to the 1.8% broad-cohort rate and the 8.9% EIN-based rate.
A rate is a question about who was counted
The useful conclusion is not that one rate is the correct rate and the others are misleading. Each rate answers a different question. The broad CSP estimate asks what happens across a tax-linked startup universe that includes many activities with no employer ambition. The EIN-based comparison asks how transition differs among businesses with an additional administrative marker. The substantive-firm analysis asks about a still narrower group with capital, owner time, and founder status. The application study asks how often prospective entrants generate employer employment within a defined twelve-month window, and how their hiring expectations compare with that outcome.
These denominators should change the reader’s interpretation of the headline. A broad population can contain many businesses that are stable, flexible, supplemental, or necessity-driven, alongside a smaller group that wants to hire. A narrower sample may show a higher transition rate because it contains more of the latter group, more resources, or both. Neither result settles whether an individual owner should hire.
The first-employee question becomes tractable only after these distinctions are kept in view. No paid employees tells us the business’s current payroll state. It does not tell us the owner’s intention, income, seriousness, survival status, or capacity to grow. The next step is to follow the businesses that do cross the boundary: when the transition occurs, how many remain active, and what separates a first employee from a broader pattern of job creation.