Starting a business is often portrayed as a test of persistence. Experienced entrepreneurs are assumed to have an advantage because they understand the process, know what can go wrong and have already learned how to bring an idea to market. New research suggests that experience can also expose founders to a different kind of vulnerability: larger ambitions require more money, and financing shortfalls may therefore weigh more heavily on those who have started businesses before.
A study published on 7 October 2026 in Small Business Economics analysed 10,883 responses to three years of a nationally representative US entrepreneurship survey. Among respondents who reported how much startup capital they had deployed, experienced founders had a weighted median of $6,000, compared with $2,500 for first-time founders. That is 2.4 times as much capital.
Yet experience did not consistently protect people from withdrawing before a business was formed. As financial disadvantages accumulated, the statistical advantage associated with previous startup experience became smaller and eventually could no longer be distinguished from zero. The finding challenges the simple idea that knowing how to start a company necessarily makes a founder more resilient to a shortage of resources.
Importantly, this is an observational study, not an experiment in which entrepreneurs were randomly given or denied financing. It shows a pattern consistent with financial constraints becoming binding for more capital-intensive plans. It cannot establish that those constraints caused any particular person to stop.
The overlooked stage before a company exists
Business research often focuses on two visible events: the creation of a company and its eventual survival or failure. A less visible group never reaches either milestone. These are people who consider a venture, investigate customers, explore finance, draft plans or begin organising a business, but then put the idea aside before an operating firm exists.
Economist Scott Shane of Case Western Reserve University examined this earlier stage, sometimes called the pre-entry margin. His paper, “Knowing when to quit: experience, binding financial constraints, and withdrawal from new venture creation”, asks whether financial limitations help explain why people discontinue an attempted startup before launch.
The distinction matters. Someone who withdraws from a startup attempt has not necessarily run a failed business. They may have decided to wait, changed their mind, or concluded that the intended venture was not feasible with available resources. Some will try again later. Counting all such decisions as business failures would misunderstand both the data and the decision being studied.
The study draws on a classic economic model of entrepreneurial choice developed by David Evans and Boyan Jovanovic in 1989. Its central idea is straightforward: the money a founder can realistically deploy may be lower than the amount required to build the venture they want. When that financing ceiling is below the venture’s capital requirement, it becomes a binding constraint. When the planned business is affordable, the same apparent financial weakness may have little practical effect.
That produces a counterintuitive prediction. A first-time founder planning a small, low-cost operation might be able to proceed despite limited savings. An experienced founder planning a larger venture could face a serious barrier with exactly the same savings, credit record and banking access.
How the researchers examined startup withdrawal
Shane used the Entrepreneurship in the Population, or EPOP, survey conducted by NORC at the University of Chicago. The analysis covered the 2023, 2024 and 2025 survey waves, using survey weights and statistical methods intended to account for the complex sampling design.
The pooled dataset contained 3,980 responses from people still pursuing a startup and 6,903 responses from people who had considered or pursued one and then stopped. Together these amounted to 10,883 survey observations. They should not be interpreted as 10,883 necessarily distinct individuals because the public files do not allow the researcher to identify whether someone appeared in multiple years.
Previous startup experience was measured by whether respondents had started a business before. To capture financial weakness that plausibly existed before the current attempt, the analysis counted three conditions: insufficient savings, a poor credit history and no established relationship with a bank or lender. Each person could have a score from zero to three.
This count is not a direct measure of the money a founder needs or can borrow. Rather, it is an indicator of financial circumstances that might limit the amount available. The researcher also considered two time-related barriers, including difficulty balancing work and family responsibilities and difficulty finding time.
The models accounted for demographic and economic factors including age, education, household income, employment status, gender, region, debt difficulties, family business background and related work experience. They also considered how far respondents had progressed through a list of 26 possible startup activities, such as market research, networking, planning and seeking finance.
Survey-weighted logistic regression was used to examine the odds of stopping, while capital deployment was analysed separately. The researcher tested whether financial constraints were more strongly associated with withdrawal among experienced founders, whether that difference varied with the number of constraints and whether alternative explanations fitted the data.
Experienced founders reported more than double the capital
The clearest descriptive result concerned the scale of the ventures being attempted. Among 6,060 respondents with valid capital amounts, the weighted median startup capital already deployed was $6,000 for experienced founders and $2,500 for novices. The gap was visible in each of the three survey years.
After adjustment for measured characteristics and whether a respondent had stopped, experienced founders were estimated to have deployed approximately 93% more capital than first-time founders. The model’s log-capital coefficient was 0.660, with a 95% confidence interval of 0.392 to 0.928 and a p-value below 0.001. Adding the number of startup steps completed reduced the estimated difference to approximately 37%, but it remained statistically detectable.
These figures answer different questions. The $6,000 and $2,500 medians describe the middle of each group’s reported capital distribution, while the adjusted percentage estimates compare groups after accounting for other measured differences. Neither number means every experienced founder needs 2.4 times as much funding as every novice.
The capital result also comes with an important caveat. People who had withdrawn were more likely to have missing capital data, so the subset reporting amounts may not perfectly represent all respondents. The researcher explicitly limits the capital comparison to those who provided usable figures.
Why the same financial problem can matter more to an experienced founder
At first glance, the financial constraint counts looked almost identical across the two main groups. Respondents who stopped averaged 0.87 of the three financial disadvantages, while those still pursuing a venture averaged 0.86. Looking only at those averages could suggest that financial conditions have little to do with withdrawal.
The more informative result emerged when the analysis considered startup experience. In the fully adjusted model, each additional financial disadvantage was associated with approximately 16% higher odds of stopping before the experience interaction was added. Previous startup experience, considered on its own in the baseline model, was associated with roughly half the odds of stopping.
However, the interaction between experience and the financial constraint count had an odds ratio of 1.274, with a 95% confidence interval of 1.027 to 1.580 and p = 0.027. This means that the association between each additional financial disadvantage and stopping was stronger for people with previous startup experience than for first-time founders.
In the interaction model, each extra constraint was associated with roughly 10% higher stopping odds among novices, an estimate that was only marginally significant. Among experienced founders, the corresponding combined association was about 40% higher odds per additional constraint.
Odds are not the same as probabilities. A 40% increase in odds does not mean that 40% more founders definitely stopped, and it cannot be converted into a universal percentage-point increase without knowing the starting probability. The finding describes a conditional statistical relationship across survey responses.
The interaction pointed in the same direction in each survey year, but was statistically significant only in 2023. The pooled estimate remained significant when the number of completed startup steps was removed from the model, which reduces concern that the result depended entirely on controlling for progress made during the attempt.
Three financial disadvantages narrowed the experience advantage
The most striking comparison examined how the relationship changed as the three financial disadvantages accumulated. With one measured constraint, experienced founders had 0.475 times the stopping odds of novices, after adjustment. The 95% confidence interval was 0.336 to 0.671, indicating a clear statistical advantage.
With two constraints, the estimated ratio rose to 0.656, with a 95% confidence interval of 0.419 to 1.025. The result was marginal rather than conventionally statistically significant. With all three constraints, the estimated ratio was 1.056, with a wide 95% confidence interval of 0.539 to 2.071 and p = 0.873.
In plain language, the data could no longer establish that experienced founders were more likely to keep going when savings, credit history and banking relationships were all problematic. This does not prove that their experience had become worthless or that the two groups had exactly equal chances of withdrawal. The confidence interval was too wide for that conclusion.
The specific interaction for the most severe constraint category was statistically detectable, with an odds ratio of 2.416 and p = 0.018. But the study did not establish a smooth, statistically distinct increase across every severity level. A formal comparison of the three interaction coefficients gave p = 0.105. The stronger conclusion is that the experience advantage weakened under severe financial pressure, not that each extra disadvantage produces a precisely predictable step in a universal progression.
People who stopped frequently pointed to money
The survey also asked people who withdrew why they had stopped. Financial resources were among the three most frequently mentioned reasons. Each additional financial disadvantage was associated with 2.271 times the odds of citing insufficient financial resources as a reason for stopping.
When respondents identified their primary reason, the odds of naming financial resources rose by a factor of 1.856 per additional constraint, with a 95% confidence interval of 1.591 to 2.166. This association was visible separately in 2023, 2024 and 2025, with year-specific odds ratios of 2.091, 1.770 and 1.822.
The pattern was especially informative among respondents who said they had been very or extremely interested in starting a business. In that group, containing 2,307 withdrawn respondents, financial constraints and their interaction with experience remained statistically associated with stopping. Among the 1,445 respondents who had been not at all or only slightly interested, those associations were not statistically detectable.
This supports the idea that the result is not merely about casual curiosity fading. Still, a retrospective survey cannot fully separate a genuine financing barrier from a person’s later interpretation of why an idea was abandoned.
What happened when founders sought outside funding?
Another analysis looked at people who approached outside funding sources, including banks, credit cards, investors, grants and family or friends. Respondents were classed as funding-rationed if at least one source rejected a request or provided less than requested.
In the pooled sample of funding seekers, rationing was associated with higher stopping odds, with an odds ratio of 1.757, a 95% confidence interval of 1.202 to 2.567 and p = 0.004. However, this result did not hold reliably when the years were examined separately. The 2025 funding-seeker sample was small, and the combined estimate for the two larger years was not statistically significant.
The researcher therefore treated this as suggestive rather than firm evidence. It would be misleading to claim that a rejected funding application was shown to cause withdrawal, or that a 76% increase in stopping odds is a dependable effect across all survey years.
Tests of competing explanations also mattered. The researcher examined whether experienced founders might stop because they had better employment opportunities or because they had lost interest after exhausting themselves. Interactions involving employment, income, new jobs, promotions and loss of focus did not explain away the central experience-by-finance pattern. These null findings weaken those particular alternatives, but they cannot eliminate every unmeasured explanation.
What founders, lenders and policymakers can learn
The practical message is not that experienced founders should always persist or that novices should abandon ambitious ideas. It is that financial readiness is relative to the business being attempted. A founder who has successfully launched a company before may be pursuing a venture with greater staffing, inventory, equipment or working-capital requirements than a first-time entrepreneur planning a smaller operation.
For aspiring founders, a more useful question than “Do I have enough savings?” may be “How much capital does this specific plan require before it can generate sustainable cash flow, and how much of that amount is actually accessible?” A simple cash-flow forecast, staged investment plan and realistic assessment of borrowing capacity can expose gaps before they become expensive commitments. Those are practical interpretations, not interventions tested by the study.
For banks and enterprise-support organisations, the findings suggest why a generic measure of personal financial weakness may be a poor guide to which entrepreneurs face the most binding barriers. The interaction between financing capacity and venture scale may matter more than either factor alone. Any proposal to expand credit, guarantees or grants would still require independent evidence about default risk, business quality and the effects of the policy.
For educators and startup advisers, the work challenges the habit of equating persistence with entrepreneurial quality. Deciding not to proceed when the financing available cannot support a viable plan can be a rational business judgment. The data do not show whether the withdrawn ideas would have succeeded if financed, but they do show why treating every decision to stop as a failure of motivation is too simplistic.
Research Today has also covered how entrepreneurial and market orientation related to performance in 293 minority-owned enterprises. That research concerned the performance of businesses, whereas the new study concerns the earlier decision of whether a business attempt continues at all. The two stages should not be confused.
Important limitations before applying the findings elsewhere
The evidence comes from US survey respondents observed at different points in time, not from a longitudinal experiment following the same people from idea to launch. A person classified as continuing might stop the next month, while someone classified as withdrawn might restart later. The study cannot establish the duration or permanence of those decisions.
The pooled 10,883 observations may include repeated appearances by the same individuals. The author performed a sensitivity calculation suggesting that the main interaction would survive an assumed 20% duplication rate with perfectly correlated responses, but the actual amount of duplication cannot be verified from the public data.
Several key measures are imperfect. The three financial conditions were reported during or after the startup attempt, so the study cannot confirm that each existed beforehand. Savings could have been depleted by the attempt itself. Desired capital, actual borrowing limits and the quality of the proposed venture were not directly observed. Missing capital responses were also unevenly distributed across groups.
Because the analysis is observational, it cannot prove that a financial constraint caused someone to withdraw or that changing a person’s credit access would make them launch a viable company. It also does not establish whether withdrawal was ultimately a good or bad decision for the person concerned.
Finally, the results describe the United States during 2023 to 2025. Financing systems, informal business activity, credit markets and startup costs differ internationally. The broad economic logic may be relevant to countries such as South Africa, but the numerical estimates should not be transferred to another market without local evidence.
The bigger lesson: experience changes the size of the problem
Entrepreneurial experience usually brings knowledge, networks and a better understanding of how a business operates. The new analysis does not contradict those benefits. Instead, it suggests that experience can change the scale of what a founder attempts, which changes how restrictive the same financial circumstances become.
That distinction helps explain an apparent puzzle in entrepreneurship research: personal wealth can appear only weakly related to starting a business when people with very different capital needs are averaged together. A financing ceiling matters little when a venture fits below it and much more when the plan exceeds it.
The study’s contribution is therefore not a simple message to persevere or quit. It is a more precise explanation of why experienced founders may be especially sensitive to severe financial barriers before a business has even opened its doors.
Source Information
Study: Shane, S. “Knowing when to quit: experience, binding financial constraints, and withdrawal from new venture creation.”
Journal: Small Business Economics.
Published: 7 October 2026.
DOI: 10.1007/s11187-026-01281-w.
Study design: Observational analysis of nationally representative US Entrepreneurship in the Population (EPOP) survey responses from 2023 to 2025, comparing 3,980 responses from continuing nascent entrepreneurs with 6,903 from respondents who withdrew before business formation. Survey-weighted logistic and capital regressions examined previous startup experience, three financial disadvantages, capital deployed and stated reasons for stopping.
