The First Year Has Two Hiring Margins
The Norwegian evidence looks closely at the moment when a new venture becomes an employer. Using register data from 2004 through 2013, Astrid Kunze and Bram Timmermans can observe who owns and leads a newly founded private company, when an employment contract begins, and how much work that contract contains. This chapter uses that unusually precise sequence to separate four questions that are often collapsed into “hiring”: whether a venture hires, how many people it hires, who those people are, and whether the work is full-time or part-time.
What the register sequence can—and cannot—show
The study begins with the problem of identifying a genuine new venture. A first registration date can be misleading: a new organization number may reflect a reorganization, a change of ownership, a merger, a subsidiary, or a spinoff rather than a new operating company. The researchers restrict the sample to private limited corporations, remove foreign-owned or already controlled firms, exclude ventures with employees registered before foundation, remove implausibly large founding-year employers, and impose activity and sector restrictions.
These filters leave a pre-CEO-linking pool of 59,910 newly registered ventures. After requiring one identifiable CEO and usable characteristics, the analytical sample contains 51,354 firms. The researchers then use employer–employee records to observe employment spells that begin during the first twelve months after registration. The final data identify 46,023 employees hired by 14,126 ventures during that first year.
The denominator matters. Only 27.5% of the ventures hire at least one employee during the period. Across all firms, the average is 0.90 employees and 0.70 full-time equivalents. Conditional on hiring, the averages rise to 3.26 employees and 2.56 FTE. A statement about the probability of making a first hire is therefore describing a minority of firms crossing a threshold, not the average employment scale of every new venture.
The timing is one of the study’s main strengths. The founding team is identified before the first-year employment spells, so a person hired later cannot normally determine who was recorded as the CEO or owner at the founding date. This ordering reduces one obvious reverse-causality problem. It does not make the founder categories randomly assigned. A female CEO may lead a different kind of venture, have different resources or networks, or be selected by a board with different growth ambitions. The authors interpret their regressions as descriptive correlations and control for many observable differences rather than claiming a clean causal effect.
The role definitions make the Norwegian design more precise than a simple male-versus-female founder comparison. The reference group is a male CEO who is also an owner. The other main categories are a female owner-CEO, a female non-owner CEO, and a male non-owner CEO. The models also account for founding-year and four-digit industry differences, 91 economic regions, financial resources measured through equity and debt, founding-team size, family-firm status, CEO education, age, citizenship, and women among owners and other board members.