The Real Answer: Both Learning and Selection

The subject of this series is serial and portfolio entrepreneurs and the firms linked to their ownership histories. The final question is whether their apparent advantage is learned, selected, or compounded—and what a reader can legitimately infer from two studies that measure different kinds of success. The most defensible answer is conditional: both learning and initial difference matter, but neither source can assign them a clean causal percentage or turn one type of entrepreneur into a universal winner.

Both stories are present, but neither wins alone

Start with what the evidence actually observes. In Portugal, portfolio serial entrepreneurs are a small group: about 3.5% of business owners, whose firms represent about 13.5% of businesses. Yet those firms account for roughly 20% of aggregate productivity growth, more than 23% of employment, and about 23% of job creation, alongside about 18% of job destruction. Their firms enter larger, grow faster over the observed life cycle, exit less often on average, and appear nearly three times as likely as other firms to be gazelles. This is a substantial and economically relevant pattern.

It is not, however, a single “serial entrepreneur effect.” The percentages use different denominators and describe different margins. The 3.5% refers to owners; 13.5% refers to firms; employment and job creation refer to economic activity. A group can employ more people than its share of firms if its businesses are larger. It can account for a large share of job creation without creating jobs at a higher gross rate if it destroys fewer jobs. It can contribute to productivity growth through changes within firms, reallocation, and the entry or exit of firms whose productivity differs from the surrounding economy. Contribution is an aggregate fact, not a measure of how much learning occurred inside each owner.

The life-cycle pattern points in two directions at once. Portfolio firms start more than 1.5 times larger than other firms and grow by about 122% between startup and age ten, compared with about 86% for other businesses. Starting larger is already a difference before later venture experience can explain anything. Continuing to grow faster may be consistent with accumulated knowledge or better execution, but it may also reflect the kinds of opportunities selected, financial room available, or the ability to keep a firm alive. The observed path contains both a starting margin and a subsequent margin.

The first-versus-subsequent comparison makes the distinction sharper. Later firms of portfolio owners are about a quarter larger in an unconditional comparison, which is compatible with learning and favorable post-entry accumulation. Yet first firms of those same owners already outperform other firms. After controls, subsequent firms remain larger but tend to exit more frequently and grow somewhat more slowly than first firms. If repeated activity alone created the advantage, the first firms should not already display it, and later firms should not change direction across outcomes. The results instead describe a history in which selection precedes learning and feedback can reinforce both.

Selection should not be reduced to a claim that successful owners possess an innate essence. It includes observable education, past wages as a proxy for ability, managerial experience, financial position, and potentially unobserved networks, institutions, motivation, or opportunity access. It also includes the fact that an owner has to survive and proceed far enough to own another business. These are different kinds of selection, but they all challenge the idea that the label “portfolio entrepreneur” is randomly distributed among otherwise identical people.

Learning should not be reduced to a classroom skill either. It can include better judgment, routines, knowledge of markets, relationships, emotional regulation, and the ability to interpret feedback. It can also include ex-post accumulation: an earlier venture may leave an owner with more net worth, better credit, or stronger credibility. A subsequent firm can therefore benefit from the first venture even when the transferable input is financial capacity rather than a newly acquired operating technique. In real ownership histories, knowledge and resources can move together.

The Polish study adds a different kind of evidence. Among 326 micro- and small-enterprise entrepreneurs in Poland, overall entrepreneurial success did not differ significantly among novice, serial, and portfolio groups, and subjective firm performance did not differ significantly either. Opportunity identification showed only a weak, small tendency for the experienced groups to score higher. Opportunity exploitation did differ: serial entrepreneurs had the highest mean rank, while portfolio entrepreneurs had the lowest. The pattern says that type can be associated with a particular stage of action without producing a universal ranking of success.

That result is not a refutation of the Portuguese administrative findings. The studies use different populations, countries, time periods, and units of observation. The Portuguese analysis follows owners, firms, and workers across a long administrative panel and measures realized employment, size, growth, productivity, exit, and aggregate contributions. The Polish study uses a 2024 online survey and asks entrepreneurs to report on opportunity processes and perceived firm performance, alongside psychological characteristics. One can find a realized firm-level premium in one design and no difference in a combined subjective score in the other without the studies measuring the same outcome and disagreeing about it.

The contrast is instructive because it answers the title at two levels.