From Firm Advantage to Economic Footprint
Once the ownership categories are fixed, the next question is not whether portfolio serial entrepreneurs sound more capable. It is how their firms move through time: what they bring at entry, what happens while they survive, how quickly they grow, and how their gains and losses add up across the economy. Following that life cycle turns a vague claim about “better entrepreneurs” into a set of observable transitions, each of which can be compared with other firms and interpreted with its own caveat.
The life cycle starts before growth
The Portuguese evidence begins with a familiar pattern known as “up-or-out.” Young firms are more likely to shut down, while the young firms that survive can grow rapidly. In the data, about 10% of the youngest firms close in an average year; by age six, the share is about 8%, while surviving startups almost double in size over that period. The important point is not that every young firm follows this path. It is that an average firm population contains two distinct transitions: some businesses leave, and others expand.
Portfolio serial-entrepreneur firms enter this process from a different starting point. They begin 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. Their exit rate is lower on average and essentially flat over the life cycle, whereas other firms exit more often. If we looked only at the endpoint, we might call this a growth advantage. Following the path shows that the difference begins at entry, continues among survivors, and includes a lower probability of leaving the observed population.
That sequence changes the interpretation. A larger starting firm may have more capital, a stronger network, better information, or a different opportunity. It may also be the result of an owner choosing a larger initial scale because they can manage several ventures or access resources that a novice owner cannot. Faster subsequent growth could reflect better decisions, learning, market selection, or the ability to finance expansion. Lower exit could signal resilience, but it can also partly reflect the fact that a business must survive before its owner can be recognized as a portfolio serial entrepreneur.
A life-cycle comparison must therefore keep states and transitions separate. At entry, a firm has a starting size and an owner classification. If it survives, it enters the continuing-firm state, where employment can rise or fall and productivity can change. If it grows quickly enough under a sustained definition, it may enter the high-growth or “gazelle” group. If it shuts down, the firm leaves the active population. The averages observed at age five or ten are shaped by both the firms that remained and the firms that disappeared before the measurement point.
The diagram is not a claim that the portfolio path causes the growth path. It is a reminder that the same label can be connected to different points in the sequence. Portfolio status is identified through ownership history, while survival and growth are outcomes observed after entry. If the classification requires an owner to have kept an earlier business alive and then own another business, the portfolio group is selected on a transition that other firms have not yet completed.
This is the first boundary on the life-cycle result. It does not make the result meaningless. It means the comparison answers a specific question: how do firms associated with owners who reach portfolio status behave, on average, relative to the other firms in the Portuguese data? It does not answer the counterfactual question of how a randomly selected novice’s firm would perform if that novice were simply given the label “portfolio entrepreneur.” The label is not a treatment assigned at birth.
The reported counterfactual makes the aggregate importance concrete. Ignoring general-equilibrium effects, the study estimates that without portfolio serial-entrepreneur firms, ten-year-old firms would on average be more than 12% smaller and average exit rates would be about 0.4 percentage points higher. This is a description of how the observed economy would look under a restricted comparison, not a forecast of a policy that removed or created portfolio firms. It shows the footprint associated with the group while leaving open why the footprint exists.
The distinction between starting scale and later growth is especially important for the question of learning. If a portfolio firm is already larger at entry, the later size gap is not entirely a reward for lessons learned in an earlier venture. If it then grows faster while continuing to survive, learning may be part of the story, but the data must still distinguish growth of continuing firms from the composition effect created when weaker firms exit. Averages can combine these mechanisms unless the life cycle is read one transition at a time.
The same caution applies to the word “survival.” A lower exit rate can represent better management, stronger financing, a more favorable market, or an owner’s willingness and ability to keep supporting a firm. It can also arise because the comparison group has been filtered by the very survival needed to become a portfolio owner. In other words, portfolio firms may be less likely to exit because of characteristics associated with portfolio ownership, because their owners respond better to shocks, or partly because only survivors enter the category. The descriptive pattern cannot choose among those explanations by itself.
Still, the pattern is not a trivial artifact. The Portuguese study observes a broad difference in starting size, life-cycle growth, exit, and employment, and reports similar conclusions under robustness checks described by the authors. The appropriate response is neither to dismiss all of it as selection nor to treat it as proof of superior entrepreneurial talent. The useful response is to ask which margins carry the difference and which of those margins matter for the aggregate economy.
From individual firms to aggregate flow
One margin is employment. Portfolio serial-entrepreneur firms represent less than 14% of firms but employ more than 23% of the workforce in the Portuguese data. That gap follows naturally from their larger average size. It also explains why an owner-level prevalence statistic cannot be used as a proxy for economic importance. A small number of owners can control a set of firms whose employment footprint is much larger than their share of people.
Another margin is the flow of jobs. Portfolio firms are responsible for about 23% of job creation and about 18% of job destruction. Their net contribution is therefore not explained by creating jobs at a dramatically higher gross rate. Among continuing firms, gross job-creation rates are comparable—about 29% for other firms and 28% for portfolio firms—but portfolio firms destroy fewer jobs, about 13% compared with 17% among other businesses. Their net job-creation advantage comes primarily from the lower destruction margin.
That distinction is easy to lose in a headline. “This group creates 23% of jobs” sounds like a statement about expansion alone. The fuller account is that the group is larger, employs more people, creates a large share of jobs, and destroys a smaller share of jobs than the comparison group. Job creation and job destruction are separate flows. A firm can add workers in one period and cut workers in another; an economy can display high reallocation even when net employment changes little. Understanding the footprint requires following both directions.
The study also reports greater excess job reallocation among other firms than among portfolio firms: about 34% compared with 26%. Excess reallocation is the portion of total job movement beyond realized net employment growth. In this comparison, the portfolio firms’ steadier employment path is not the same as inactivity. They still create and destroy jobs, but their lower destruction means less churn relative to the other group. The result points toward a particular form of dynamism—growth with fewer losses—rather than simply more movement in every direction.
Productivity requires a similar decomposition. Aggregate productivity growth is not an average of owner characteristics. It reflects what happens inside continuing firms, how employment and market weight are reallocated across firms, and what entering and exiting firms add or remove relative to the existing economy. A portfolio group can contribute through its own productivity changes, through expansion that shifts weight toward its firms, and through the fact that its entering or exiting firms differ from the firms around them. Saying that portfolio firms account for about 20% of aggregate productivity growth identifies their aggregate contribution; it does not say that 20% of each firm’s productivity came from serial experience.
This distinction between contribution and cause is central. An aggregate decomposition tells us which group is associated with a share of a change after the economy’s firm-level movements have been organized into components. It does not by itself identify the owner-level intervention that produced the change. A firm may have grown because its owner learned, because it entered a favorable market, because it had more financing, or because the aggregate calculation assigns more weight to its expansion. The decomposition makes the footprint visible; other comparisons are needed to explain the mechanism.
High-growth firms make the concentration even clearer. Under the Eurostat–OECD definition used in the study, a gazelle is a business up to five years old that reaches at least ten employees at some point and achieves average annualized growth of at least 20% over a three-year period. About 9% of Portuguese businesses meet this definition, yet they account for almost a third of aggregate employment and more than 40% of job creation. A rare tail of firms can therefore dominate flows that look broad when measured across the whole economy.
Portfolio owners are overrepresented in that tail. About 29% of all high-growth firms are owned by portfolio serial entrepreneurs, even though portfolio firms account for less than 14% of all firms. Under the study’s comparison, portfolio entrepreneurs are nearly three times as likely as other owners to own a gazelle. This is a powerful clue about selection into high-growth activity, but it is not yet evidence that portfolio ownership caused the high growth. The group may contain owners with resources or capabilities that make high-growth entry more likely.
Being more likely to own a gazelle is also different from being better once a firm is already a gazelle. Among high-growth firms, the portfolio-owned businesses are only slightly larger and less likely to shut down; their growth and productivity are not statistically different from those of other gazelles. The strongest difference lies in reaching the high-growth category, not in winning every comparison after entry into it. That distinction prevents a common error: using a group’s representation among exceptional firms as proof that every member of the group outperforms every comparable firm.
The gazelle definition itself illustrates how outcomes and classifications can interact. A firm becomes a gazelle because it displays a particular age, employment, and growth history. The category is useful for studying where job creation is concentrated, but it is outcome-linked. Comparing portfolio representation among gazelles is therefore partly a study of who reaches a selected tail. It cannot cleanly separate an owner’s initial characteristics from the growth process that eventually satisfies the definition.
This does not reduce gazelles to statistical tricks. The high-growth firms employ people and create jobs in the observed economy; their contribution is real. It does mean that the interpretation must match the question. If the question is where job creation is concentrated, gazelle status is informative. If the question is whether portfolio ownership causes growth, gazelle status is an endpoint that needs to be analyzed with more careful comparison and timing.
The same logic applies to the group’s lower exit profile. The difference matters for employment and productivity because keeping a firm active preserves its workers and its weight in the aggregate economy. But it also carries the strongest mechanical selection concern: to own multiple businesses simultaneously, an owner must have reached a point at which more than one business exists and remains observable. The portfolio group is thus not a random slice of new firms at entry. It is a population that has passed through earlier transitions.
The life-cycle evidence gives us a precise description of the visible advantage. Portfolio serial-entrepreneur firms enter larger, grow faster over the observed horizon, destroy fewer jobs, employ more workers, contribute disproportionately to productivity growth, and appear more often among gazelles. These margins are economically meaningful and connected. Yet they do not all point to the same underlying mechanism, and they do not establish that repeated ownership alone produced them.
That is why the next explanatory test has to move inside the portfolio group. If later ventures are larger than earlier ventures of the same owners, the pattern may reveal learning or accumulated resources. If first ventures already outperform other firms, it points toward selection before the second venture. The aggregate footprint tells us where the difference matters; the first-versus-subsequent comparison asks what kind of history could have produced it.