The First Job Is Also a Founder Decision
This series examines newly founded ventures, the founders and founding teams behind them, and the early employees who enter those organizations. Its central question is not simply whether a startup hires, but whether a founder’s background helps shape who receives an opportunity, how much work that opportunity contains, and whether the resulting workforce composition persists. That question matters because a small number of founders can influence employment for many people. The evidence can support a careful judgment about that relationship—but not the claim that similarity between founders and workers is automatically causal, intentional, or inevitable.
Startups are not neutral job gateways
A new venture is often described as a bundle of ideas, capital, and risk. It is also an employer, sometimes before it has a settled product or a formal human-resources process. The first people brought into the organization help turn an idea into a working operation. They may write the first software, meet the first customers, establish routines, or determine what kinds of work the venture is equipped to undertake. In that sense, the founding organization is a labor-market gateway: it decides who gets access to an uncertain but potentially valuable early job.
That gateway matters because access to entrepreneurship is uneven. The U.S. study by Josh Feng and Xavier Jaravel begins from the observation that young firms are important sources of net job creation while women, racial and ethnic minorities, and people with less formal education are underrepresented among entrepreneurs. The Norwegian study by Astrid Kunze and Bram Timmermans starts from a related gender gap and asks what happens at a less frequently examined point: the hiring process immediately after a new venture is formed.
The question is therefore larger than whether women or minority founders hire “more” or “less.” If the people who create organizations are drawn disproportionately from some backgrounds, and if those organizations are more likely to recruit workers from similar backgrounds, then unequal access to entrepreneurship can become unequal access to employment. A worker may be affected by the composition of the founder pool even when they never plan to become an entrepreneur themselves.
This does not mean that founders simply reproduce their own identities in every decision. A founder may hire someone with similar experience because that person has relevant skills, because they share a professional network, because the work requires trust under severe uncertainty, or because the organization’s production process benefits from particular forms of complementarity. A worker may also prefer, accept, or seek out a firm led by someone with a similar background. The evidence in these studies helps locate the pattern. It does not, by itself, settle the motive.
The practical implication is that early hiring should be read as an organizational event, not only as a private choice between an employer and an applicant. The choice establishes a workforce, and the workforce can influence later recruiting, retention, and the organization’s capacity to grow.
The timing matters. In a mature company, a vacancy may come with a settled job description, a known reporting line, and established procedures for evaluating candidates. A new venture often has none of these. The founder may still be deciding which tasks deserve a dedicated role, which capabilities are urgent, and how much uncertainty a new employee can absorb. Networks, referrals, shared experience, and confidence in a candidate can carry unusual weight in that environment. Those facts make founder background potentially consequential, but they also make it difficult to infer a single motive from an observed match.
The worker’s side of the relationship matters as well. An early startup job can offer access to responsibility and rapid learning, but it can also involve unstable hours, incomplete support, and uncertain survival. People who are already well connected to a founder may hear about an opening sooner. People who have fewer alternatives may accept a role or contract that a better-positioned worker would decline. A composition gap can therefore emerge from employer choices, worker choices, or the interaction between them. The studies’ value is to show that the relationship is systematic enough to investigate; their limits remind us not to turn a pattern into a motive without direct evidence.
This is also why the unit of analysis matters. The relevant subject is not a founder’s personal identity detached from an organization, and it is not a worker’s demographic category detached from a job. It is the venture as a place where leadership, resources, production, search, and employment meet. The rest of the series keeps returning to that organizational level: what the founder measure captures, what the employee measure captures, and which consequences can be observed as the venture adds people.
Four questions hidden inside “hiring”
The word hiring compresses several different outcomes. Keeping them separate is essential, especially because the two studies measure different parts of the process.
The first is the extensive margin: does a venture hire at least one employee? This is the threshold between an organization that remains only founder-operated and one that begins to employ someone else. It is a consequential step, but it says nothing about the size or terms of the job created.
The second is the intensive margin: how many people does the venture hire after crossing that threshold? A firm that hires one person and a firm that hires ten both count as hiring, even though their employment contributions differ greatly. Researchers may measure this in headcount or in full-time equivalents, which adjust for the number of hours represented by the contracts.
The third is employee composition: who is in the workforce? The U.S. paper compares a founder share with the share of workers who have the corresponding demographic characteristic. If a firm’s founders are more likely to be women and its employees are also more likely to be women, that is a composition relationship. It is not the same as saying that a female founder hired each female employee directly, or that every later decision followed the same rule.
The fourth is contract intensity: how much work is offered to the people who are hired. Two ventures can recruit the same number of women but offer very different amounts of paid work. A part-time contract, a full-time-equivalent measure, and the identity of the employee answer related but distinct questions about the opportunity.
A final distinction concerns time. A workforce can remain compositionally similar to its founders because later hires continue to resemble them, because initial employees stay, or because both processes operate together. Persistence is a fact about the workforce over time; it is not proof that every later hiring decision repeats the founding decision.
These distinctions prevent a common interpretive error. A venture can be more likely to make at least one hire but create fewer full-time-equivalent jobs. It can recruit a larger share of women while offering those hires fewer hours. It can show a strong founder–worker association several years later even after the association among new hires has weakened, because employees who arrived early remain. “Who hires whom” is therefore a family of questions, not a single statistic.
Two lenses on one labor-market relationship
The U.S. evidence supplies breadth. Its primary source is the Census Bureau’s Longitudinal Employer–Household Dynamics data, which combine administrative wage records with demographic, industry, and location information. In the main balanced sample, the researchers follow approximately 825,000 new firms from founding year through year five, retaining firms with non-founder employees in every year of that window. Because the data do not directly identify founders, the study uses the top three earners in a company’s first year as an operational founder measure. That makes the sample unusually broad, but it also means that “founder” is a carefully chosen proxy rather than a perfect observation.
A linked LinkedIn–Crunchbase sample provides a different advantage. Founder titles are visible, and work histories include occupations and seniority, allowing the researchers to ask whether the pattern appears among engineers, scientists, senior workers, or highly paid employees. That sample is narrower and more selective: LinkedIn participation is not universal, its demographic categories are inferred from names, and the firms are tilted toward venture-backed and technology startups. It is useful precisely because it measures different aspects of the relationship, not because it eliminates the limits of the administrative sample.
The Norwegian study supplies depth at the beginning of the sequence. Its register data distinguish owners, CEOs, board members, and employees, and record the start date of employment contracts. The authors identify newly founded private limited corporations, remove apparent reorganizations and other cases that could masquerade as new ventures, and observe hiring during the first twelve months. This allows the study to distinguish female owner-CEOs from female non-owner CEOs and to measure not only the number of employees but also full-time-equivalent and part-time work.
These lenses should be combined carefully. The U.S. studies cover multiple demographic dimensions, compare firms over longer periods, use wage data, and develop a model of aggregate implications. The Norwegian study focuses on gender, a different institutional setting, and a much tighter first-year window. Their samples, periods, founder definitions, and outcomes are not interchangeable. A pattern that appears in both settings is informative, but it is not a pooled estimate of one universal effect.
The same caution applies to the evidence standard. A controlled comparison can show that founder composition is associated with worker composition among otherwise similar observed firms. A wage decomposition can make a demand-side interpretation more plausible. A calibrated model can show how such a pattern might affect equilibrium wages. None of those steps automatically becomes a causal estimate of a policy or a complete account of human motivation.
The distinction is especially important when the two studies are read together. The U.S. evidence spans firms founded from the 1990s into the 2010s and follows workforce composition well beyond the first year, but it cannot directly observe every founding relationship or every non-wage term of employment. The Norwegian evidence sees the founding-to-hiring sequence much more closely, but only in one country, one institutional setting, and the first year of operation. Agreement on a broad direction is therefore a reason to take the question seriously, not permission to erase the design differences.
The solid path in the diagram captures the relationship the studies can observe or estimate: founders establish an organization, early employment follows, and workforce composition or contract intensity may persist. The dashed path marks what remains less directly observed—whether the relationship reflects employer demand, worker selection, referrals, production complementarities, financial constraints, or some combination.
That is the useful starting model for the evidence that follows. The subject is not founder identity as a symbolic attribute. It is the way a founder-shaped organization opens, concentrates, or changes access to early work. The next question is how the U.S. data measure that relationship and what remains after firms are compared within industry, geography, and founding cohort.