The Narrow Bridge: How a Nonemployer Becomes an Employer
The 2011–2012 startup cohorts provide a rare way to watch businesses move from revenue without payroll to paid employment. This chapter follows that lifecycle: how many businesses cross the boundary, when the first employee appears, and what happens to businesses that remain solo, close, survive, or create additional jobs. The evidence makes the bridge narrow, but it also shows why a transition rate cannot be read as a simple test of ambition.
A broad cohort mostly does not cross
The Comprehensive Startup Panel follows businesses that began operations in 2011 or 2012 and tracks the 2011 cohort through 2018 and the 2012 cohort through 2019. In total, the study observes about 8.752 million startups: 4.387 million began in 2011 and 4.365 million began in 2012. This is a broad, tax-linked universe, not a selected list of venture-backed companies or firms that had already announced a hiring plan.
At startup, 96.8% of these businesses had zero employees. By the end of the eight-year observation period, the status distribution looked like this:
- 7.801 million businesses, or 89.1%, started and ended as nonemployers.
- 159,400, or 1.8%, started as nonemployers and transitioned to employer status.
- 522,500, or 6.0%, started as employers and had no positive net employment growth over the period.
- 268,700, or 3.1%, started as employers and had positive net employment growth.
The first figure answers the headline question for this particular population and window: very few broad-cohort startups that began without paid employees were observed crossing into employer status. It does not say that 1.8% is a permanent lifetime probability for every business owner today. It says that, under the study’s definition of startup and its eight-year follow-up, this was the share classified as making the transition.
The word “ended” in this table also needs care. The four categories describe the business’s status at the end of the transition classification, but they do not by themselves distinguish an active nonemployer from one that closed before the endpoint. A business can fail without ever becoming an employer, and a business can remain active without hiring. To see those paths, survival has to be added as a separate dimension.
The broad cohort is also not a cohort of identical firms waiting at the same starting line. It includes business activities that file taxes, including some secondary work and small-scale consulting, contracting, or hobby-like activity. Many of these activities may be economically useful to the owner while having no reason to add payroll. If the denominator includes businesses that were never candidates for employer growth, the transition share will be lower than it would be among a group selected for sustained, full-time, growth-oriented operation.
That is not a weakness to be corrected by discarding the broad number. The broad number is valuable precisely because it shows what happens across the full observed business population. It tells policymakers and readers not to infer a conventional employer trajectory from the mere fact that a business has filed for taxes or recorded revenue. It also establishes why a narrower sample can produce a different answer without contradicting the broader one.
The bridge has a front-loaded entry window
The businesses that do cross the boundary tend to do so early. Among the 159,400 classified transitions, 5.6% occurred in transition year 0, 44.2% in transition year 1, 17.0% in year 2, 10.5% in year 3, and 7.6% in year 4. The cumulative share reached 49.8% after transition year 1, 66.8% after year 2, 77.3% after year 3, and 84.9% after year 4. The later shares were smaller: 5.6% in year 5, 5.3% in year 6, and 4.2% in year 7.
The indexing takes a moment to understand. Transition year 0 is the startup year; transition year 1 is the second observed year. So the largest group did not wait several years after becoming established. Nearly half of the classified transitions occurred in that second observed year, and more than three-quarters occurred by transition year 3. The pattern is front-loaded even though the study follows businesses for eight years.
The authors treat the year-zero category cautiously. A business is identified as a startup when it first has positive revenue, while employer status depends on payroll. A firm that shows payroll within the first year may have begun with employees immediately but appear in the data as a nonemployer because the measures do not line up perfectly in time. The researchers assume these cases represent genuine transitions, while noting that the largest groups are in years 1 and 2 rather than year 0. That timing concern does not remove the broader pattern, but it limits how confidently the earliest category should be interpreted.
An early transition window changes the way “not yet hiring” should be understood. If a business is going to become an employer in this cohort, the probability is concentrated in the first years of operation. A long period of solo operation is not simply a business waiting patiently for an inevitable next stage. It may indicate that the business has found a stable solo model, that demand has not become predictable enough to support a wage, that the owner lacks capital or time, or that the activity is not designed for payroll at all.
The data do not identify one universal trigger for the early pattern. Early demand discovery could reveal that a business has more work than one person can handle. The owner may have enough information to commit capital, or may encounter a customer contract that makes employment practical. Some businesses may also hire early because the legal or operational design presupposes a team. These are plausible mechanisms, but the transition-year table alone cannot tell us which one occurred in a given firm. Its contribution is to establish when the crossing is most concentrated, not why every crossing happened.
This lifecycle is not a claim that the states are equally likely or that every firm moves through them in a neat sequence. It makes two distinctions visible. First, a nonemployer can either remain without payroll or cross into employer status; both are different from closure. Second, an employer can survive without achieving positive net employment growth, while positive growth is an outcome measured over the period rather than a synonym for initial employer entry.
Crossing, surviving, and growing are different paths
When survival is included, the broad cohort separates into six groups. About 23.3% started and ended as nonemployers and survived. Another 1.3% started as nonemployers, transitioned to employers, and survived. Among businesses that started as employers, 1.3% survived without positive net job creation and 3.1% survived with positive net job creation. The remaining 66.4% started as nonemployers and did not survive, while 4.6% started as employers and did not survive.
The first implication is easy to miss: most nonemployer startups in this broad cohort did not survive the full observation period, but a substantial minority did. Continued solo operation is not one thing. It contains businesses that remain active and businesses that disappear from the relevant records. The phrase “never hired” therefore cannot be used as a proxy for “failed.” Nor can it be used as a proxy for “successful stable business” without checking the survival measure.
The survival rates sharpen the contrast. Across all startups, 28.9% survived through the observation period. Nonemployer startups had a 27.0% survival rate, compared with 48.8% for employer startups. Businesses that started and ended as nonemployers had a 26.1% survival rate, while businesses that started as nonemployers and transitioned to employers had a 68.6% survival rate.
That last comparison is striking, but it is not a controlled experiment. Businesses that eventually hire are likely to differ before the hire occurs. They may have more capital, stronger demand, a more formal organizational structure, a better customer pipeline, or an owner who can devote more hours to the business. Those same features could help survival. The high survival rate may partly reflect the kind of business that crosses the threshold, not a benefit caused by adding an employee.
The multivariate analysis points in the same direction without resolving causality. After controlling for observed firm and owner characteristics, starting as a nonemployer is negatively associated with survival relative to starting as an employer. In the full model, the coefficient is −0.4404, which corresponds to survival odds about 0.643 times those of businesses that started as employers. The estimate says that initial employment status contains information about survival prospects in this sample. It does not identify whether the absence of an employee is the cause, a marker for limited resources, or a consequence of other conditions that the controls do not fully measure.
The matched sample of businesses linked to the Survey of Business Owners has different levels because it is not the full startup universe. Its overall survival rate is 40.8%; businesses that remained nonemployers had a 36.0% survival rate, while transitioners had a 72.1% rate. The direction is similar to the full cohort, but the levels are higher. This is a useful reminder that adding detailed owner and firm information also changes who is observed. A richer sample is not automatically a more representative one.
Positive job creation is broader than the first employee
The study’s third outcome is positive net employment growth. It includes nonemployer startups that transition to employer status, but it also includes employer startups that add employees above their original base-year employment. That broader definition matters. Suppose one business starts with no paid workers and later hires one. It contributes to positive net employment growth through the transition. Another starts with two employees and later has four. It also contributes, even though employee number one was already present at startup.
Conversely, starting as an employer does not guarantee growth. In the broad cohort, 6.0% of startups began as employers but had no positive net employment growth over the observation period. An initial payroll may be enough to classify the business as an employer without being large enough, stable enough, or well matched to demand to produce additional net jobs. Employer status is a state; net growth is a change from a baseline.
The positive-growth regression reinforces the role of initial selection. Starting as a nonemployer has a large negative association with positive net employment growth in the full model, with a coefficient of −2.476 relative to starting as an employer. Corporations and partnerships show positive associations relative to sole proprietorships, most positive financial-capital categories are positive relative to no capital, and website presence is positive. These results describe patterns in the data, not a causal penalty assigned to every owner who starts alone.
The distinction between the outcomes also changes what “success” means. For a policymaker focused on job creation, the first employee is an important event because it extends the benefits of a business beyond the owner. For an owner focused on autonomy or reliable personal income, remaining a nonemployer may be a satisfactory endpoint. For a researcher studying firm dynamics, survival may matter more than a single hiring event. These judgments can coexist because they refer to different objectives.
The lifecycle evidence therefore gives a precise answer and leaves a precise question. The answer is that the transition from nonemployer to employer is rare in the broad 2011–2012 cohorts, concentrated early, and associated with much higher observed survival among the businesses that make it. The question is what distinguishes those businesses before the transition: not just whether they want to hire, but whether they have the capital, organizational form, demand, time, and operating conditions to carry the new obligation. That is where the first employee stops looking like a missing checkbox and starts looking like a threshold in how the business works.