Original Research

Gendered drivers of entrepreneurial intentions: Evidence from South Africa using Global Entrepreneurship Monitor data

Takawira M. Ndofirepi, Renier Steyn

Received: 11 Aug. 2025; Accepted: 07 Apr. 2026; Published: 01 Sept. 2026

Copyright: © 2026. The Authors. Licensee: AOSIS.
This work is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license (https://creativecommons.org/licenses/by/4.0/).

Abstract

Background: Entrepreneurship is a key driver of economic activity, yet persistent gender gaps in entrepreneurial intentions lead to unequal participation between men and women in the economy.

Aim: This study examines the role of key factors in shaping entrepreneurial intentions among men and women in South Africa.

Setting: Cross-sectional data from the 2019 Global Entrepreneurship Monitor Adult Population Survey.

Methods: Data comprised of responses from 1475 men and 1516 women. A logistic regression model was employed to test the predictive relationships between demographic, perceptual and contextual factors and entrepreneurial intentions across genders.

Results: Results indicate that age, self-efficacy and knowledge of entrepreneurs significantly predict entrepreneurial intentions for both men and women. However, fear of failure significantly deters women but has no statistically significant effect among men, highlighting a gender-specific psychological constraint.

Conclusion: The study identifies the primary drivers in the entrepreneurial ecosystem and points out where gender-sensitive interventions may be useful.

Contribution: This study provides a comprehensive examination of gendered factors influencing entrepreneurial intentions in the South African context, using a broad sample.

Keywords: entrepreneurial intentions; fear of failure; self-efficacy; social cognitive theory; South Africa; Global Entrepreneurship Monitor.

Introduction

Entrepreneurship serves as an important mechanism for poverty alleviation, inequality reduction and the economic empowerment of marginalised groups, including youth, women and rural populations (Acar & Findikli 2020; Cummings & Lopez 2022; Khursheed 2022; Ndlovu & Lefera 2024; Ogamba 2019; Omeje, Jideofor & Ugwu 2020). Despite its promise of inclusiveness, entrepreneurial activity remains unevenly distributed across genders worldwide (Agrawal et al. 2023; Angulo-Guerrero et al. 2024; Bimha et al. 2018; Gaies et al. 2023), with women continuing to be under-represented in the entrepreneurial landscape. Data from 81 economies show that women constitute only 25% of new business owners (World Bank 2024), and the Organisation for Economic Co-operation and Development (OECD) reports that women are 30–50% less likely than men to start a business, secure start-up capital, or own established firms (OECD 2023). These disparities reflect not only constraints faced by women, but also differential advantages historically afforded to men. Although men and women operate within the same macroeconomic and institutional environment, men often benefit from differential exposure to entrepreneurial networks, more permissive social norms regarding risk-taking, stronger labour market positioning and greater tolerance for uncertainty (Marlow 2020; Motoyama et al. 2021), factors that may collectively translate into higher entrepreneurial participation relative to women under similar contextual constraints.

The persistence of these gendered patterns is particularly noticeable in South Africa, a context of significant interest. Despite possessing comparatively developed financial markets and entrepreneurship support structures for the region, the country’s entrepreneurial activity remains below global averages, underscoring a notable paradox (Bowmaker-Falconer, Meyer & Samsami 2023). This underperformance persists against a backdrop of severe structural economic challenges including income inequalities (World Bank 2022), which, in turn, lead to necessity-driven entrepreneurship as a survival strategy (Bowmaker-Falconer and Herringto 2020). These conditions form the structural backdrop against which entrepreneurial intentions are formed. Despite some improvement in women’s participation, overall entrepreneurial intentions remain low, with only 10% of adults aged 16–64 years indicating an intention to start a business in 2023. These patterns raise critical questions about why start-up rates remain suppressed and how entrepreneurial intentions continue to differ between men and women under the same structural conditions.

Although several studies in South Africa have examined entrepreneurial motivations (e.g. Malebana 2015), existing evidence remains fragmented and limited in scope. Much of the literature relies on samples drawn from higher education institutions (e.g. Ndofirepi, Rambe & Yao Dzansi 2018; Shambare 2013) or predominantly urban populations (e.g. Boucher, Cullen & Calitz 2024; Urban & Moloi 2022), thereby restricting generalisability to the wider urban population. As a result, there is limited nationally representative evidence on how demographic characteristics, perceptual factors (such as self-efficacy and fear of failure) and contextual conditions (including opportunity perception and societal valuation of entrepreneurship) interact to shape entrepreneurial intentions differently for men and women in South Africa. This gap limits a more detailed understanding of gendered entrepreneurial intentions other than student- and urban-centred analyses.

By utilising South Africa’s GEM dataset, the present study addresses this gap, offering a more inclusive and contextually grounded analysis of the gendered determinants of entrepreneurial intentions. In this context, the ‘less explored’ populations refer chiefly to the wider adult population represented in GEM, groups seldom included in South African studies on gender differences in entrepreneurial intentions.

This study recognises that entrepreneurial intentions are influenced not only by demographic attributes, gender included, but also by cognitive-perceptual factors and the larger contextual environment in which individuals are embedded (Donaldson 2019; Liñán & Fayolle 2015; Ndofirepi & Steyn 2023). Prior studies have demonstrated that these determinants often differ for men and women (Arshad et al. 2016; Nikou et al. 2019). Yet little is known about how these determinants differ by gender in South Africa, highlighting the need for a gender-sensitive analysis that accounts for these multidimensional influences. Against this background, the present study investigates how demographic, perceptual and contextual determinants vary across gender using the 2019 GEM data, the latest publicly available.

The following research questions were addressed:

• How do demographic factors (age, household income) influence entrepreneurial intentions differently for men and women?
• What is the role of perceptual factors (self-efficacy, fear of failure, knowledge of entrepreneurs) in shaping gendered entrepreneurial intentions?
• How do contextual factors (entrepreneurial prestige, ease of starting a business, perceived opportunities and media coverage) affect men and women differently in their entrepreneurial aspirations?

By addressing these questions, this study aims to contribute to the literature on gender and entrepreneurship by providing evidence-based insights within the broader South African national context. The findings have important implications for national policymakers, educators and practitioners seeking to develop broad-based interventions that reduce gender-based entrepreneurial barriers across the country.

Literature review

In this section, the discussion addresses four key matters: the theoretical frameworks relevant to the study, the South African entrepreneurship context, empirical evidence on demographic, perceptual and contextual factors shaping gendered differences in entrepreneurial intentions, and the research gap these insights collectively reveal.

Theoretical perspectives

The study draws on three key theoretical perspectives to explain gender differences in entrepreneurial intentions: Bandura’s (1986) social cognitive theory (SCT), North’s (1990) institutional theory (IT) and Eagly’s (1987) gender role theory (GRT).

Social cognitive theory posits that self-efficacy – an individual’s belief in their ability to execute entrepreneurial tasks – plays a central role in shaping entrepreneurial behaviour (Krueger & Carsrud 1993). Self-efficacy influences motivation, resilience and risk-taking, traits foundational to entrepreneurial action (Bullough, Renko & Myatt 2014; Emrizal & Primadona 2023; Kunshu & Yan 2015). Previous studies consistently show that women report lower entrepreneurial self-efficacy than men, which may partly explain their lower entrepreneurial intentions (Álvarez-Huerta, Muela & Hermida 2022; Caliendo et al. 2023; Laguía et al. 2022; Mansour 2019). Conversely, men may benefit from higher exposure to entrepreneurial role models, networks and societal expectations that encourage risk-taking and venture creation, reflecting differential opportunities across gender. Central to self-efficacy is vicarious learning, the process through which individuals learn from what they observe (Steyn & Mynhardt 2008). A lack of visible female entrepreneurial role models may therefore limit women’s participation in entrepreneurial activities. Accordingly, SCT provides a useful framework for understanding how the environment shapes self-efficacy and entrepreneurial intentions in South Africa.

While SCT offers a psychological lens on individual perceptions of competence, IT provides a broader socio-structural perspective, emphasising how formal and informal institutional arrangements shape entrepreneurial activity (North 1990). Institutional theory highlights how ‘rules of engagement’ in the business environment create unequal opportunities and barriers for men and women (Castellaneta, Conti & Kacperczyk 2020; Gaweł & Toikko 2024). In South Africa, women often face institutional constraints such as limited access to funding, exclusion from business networks and underrepresentation in male-dominated sectors (Jali, Tengeh & Makoza 2023; Shambare 2013; Shava & Chinyamurindi 2022), all of which may inhibit their entrepreneurial engagement.

The GRT further proposes that societal expectations shape occupational choices for men and women (Eagly 1987), including decisions to pursue entrepreneurship. Higher levels of risk aversion and fear of failure, frequently more pronounced among women due to culturally reinforced nurturing or preserving roles, can discourage entrepreneurial entry (Humbert & Brindley 2015; Nishat & Nadeem 2022; Nyanga & Chindanya 2021). Conversely, cultural norms may incentivise men towards risk-taking and leadership roles, thereby influencing the gendered distribution of entrepreneurial intentions.

Taken together, these three theoretical frameworks provide a comprehensive foundation for analysing gendered entrepreneurial intentions. Social cognitive theory explains individual-level perceptions and motivations (Kickul et al. 2008; Mueller & Conway Dato-on 2013; Panda 2018; Song & Lu 2024; Van Ewijk & Belghiti-Mahut 2019), IT situates these within broader structural and institutional contexts, and GRT explains how societal expectations shape behavioural norms (Adikaram & Razik 2023; Basset, Bell & Kharusi 2022; Raza, Yousafzai & Saeed 2024; Shambare 2013; Shava & Chinyamurindi 2022; Strawser, Hechavarria & Passerini 2021).

The South African context

Global gendered patterns in entrepreneurship are particularly salient in the South African context. South Africa faces serious economic challenges, with high unemployment, persistent economic inequality and a dual economy comprising formal and informal sectors (Statistics South Africa 2025; World Bank 2022). According to Statistics South Africa (2025), the estimated unemployment rate stood at 31.9% at the end of 2024, while youth (aged 15–24 years) unemployment stood at 62.5% and graduate unemployment was 12.2%, reflecting structural barriers to labour market and entrepreneurial engagement (Anwana & Anwana 2020; Statistics South Africa 2025). The GEM shows that entrepreneurship in South Africa is driven more by necessity than opportunity, reflecting its role as a survival strategy (GEM 2019). Although more recent GEM data exist internally, 2019 remains the latest publicly accessible wave and allows for nationally representative analysis.

Despite comparatively robust government support (UN Trade and Development [UNCTAD] 2023), entrepreneurship performance remains low. Entrepreneurial activity and early-stage participation rates are below global averages (Bowmaker-Falconer et al. 2023; GEM 2019). Women are concentrated in informal, low-productivity sectors and face restricted access to financial resources, networks and institutional support (Jali et al. 2023; Shava & Chinyamurindi 2022). Men, by contrast, frequently enjoy higher exposure to networks, social legitimacy and structural advantages that facilitate venture creation. Social and cultural expectations also influence risk perceptions, fear of failure and occupational choices, highlighting the importance of examining gendered determinants within the national context (Etim & Iwu 2019). Although women’s participation has gradually improved, significant challenges persist. Reflecting the country’s strained economic conditions, only 10% of adults aged 16–64 intended to start a business in 2023, the lowest level in two decades.

Policymakers have become increasingly aware of the need to cultivate and support resilient, innovative and agile entrepreneurs, as reflected in the National Integrated Small Enterprise Development (NISED) Masterplan, which allocated a budget of R2.437 billion for 2024/2025, of which R2.034bn (approximately 83.5%) was directed, mainly, to the Small Enterprise Development Agency (SEDA) and the Small Enterprise Finance Agency (SEFA) (Department of Small Business Development 2024). The UNCTAD (2023) pinpoints three key areas of support: improving the regulatory and enabling environment to reduce administrative barriers and create a more business-friendly climate; strengthening entrepreneurship education and skills development to build the competencies required for successful venture creation; and expanding enterprise support and inclusion, including access to finance, technology, markets and targeted support for women, youth and other underrepresented groups. Together, these areas form the core of the country’s formal entrepreneurship support system. Nevertheless, female entrepreneurial activity remains disproportionately low, indicating persistent gendered barriers.

Existing research on the gender-entrepreneurial intentions nexus

Research shows that entrepreneurial intentions are shaped by demographic, perceptual and contextual factors, which often interact across gender.

Demographics: Male and female

Several studies have examined the relationship between demographic factors, especially gender and entrepreneurial intentions. A consistent finding across the literature is that men tend to report higher entrepreneurial intentions than women, often attributed to greater access to networks, stronger exposure to entrepreneurial role models and societal norms that legitimise risk-taking behaviour (Griswold & Palmer 2020). However, this apparent male advantage is increasingly contested. Evidence indicates that when women are exposed to similar enabling conditions, gender differences in intentions may diminish or reverse. For example, in Ugandan, female undergraduate students with prior business exposure exhibited higher entrepreneurial intentions than those without such exposure (Rukundo et al. 2025). Paradoxically, the same study reported that male students whose parents were involved in business had lower entrepreneurial intentions, suggesting that family background influences entrepreneurial intentions in complex and nonlinear ways rather than uniformly motivating young people towards entrepreneurship. Sargani et al. (2021) also identified gender-related discrepancies in entrepreneurial ambitions among undergraduate students. Further complicating the picture, some studies report no statistically significant gender differences in entrepreneurial intentions. For instance, Mamatha, Chaitra and Honnali (2024) found no statistically significant differences in entrepreneurial intentions based on gender, region or field of study, indicating contextuality and intersectional complexity of demographic influences. This suggests that gendered differences in entrepreneurial intentions are neither universal nor deterministic, but shaped by the interaction of demographic characteristics with contextual and experiential factors.

Perceptual factors and gendered entrepreneurial intentions

Perceptual factors are central to entrepreneurial intentions by influencing how individuals interpret opportunities, capabilities and risks. In entrepreneurship literature, these factors include self-efficacy, risk averseness, perceived business opportunities, entrepreneurial networks, perceived ease of doing business, media support and entrepreneurial prestige, among others (Batz Liñeiro, Romero Ochoa & Montes de la Barrera 2024; Hoogendoorn, Van der Zwan & Thurik 2017). However, the influence of these factors is neither uniform nor consistently gendered across contexts.

A substantial body of literature suggests that women typically report lower entrepreneurial self-efficacy than men, largely due to socio-cultural norms and expectations (Álvarez-Huerta et al. 2022; Mueller & Conway Dato-on 2013). Yet, these relationships are not fixed. Empirical evidence suggests that enabling environments, such as entrepreneurship education and external exposure, can significantly mitigate or eliminate these differences (Austin & Nauta 2015).

Similarly, the fear of failure is identified as a major constraint, especially among women, reinforcing the gendered nature of risk aversion (Duong & Vu 2024; Gao et al. 2024; Guelich 2022). However, emerging evidence suggests that its effect is context-dependent, with some studies showing that fear of failure affects both men and women, albeit through different psychological and social mechanisms (Gao et al. 2024). This raises questions about whether fear of failure is inherently gendered or instead amplified by structural inequalities and social expectations.

In terms of network utilisation, men tend to rely on instrumental, goal-oriented ties, and women rely on relational support networks (Liu et al. 2020; Neumeyer et al. 2019), with entrepreneurial ecosystem biases limiting women’s networking efficacy, particularly in emerging economies (Pindado et al. 2023; Vieitez-Cerdeño, Manzanera-Ruiz & Namasembe 2023).

Contextual factors and gendered entrepreneurial aspirations

Gendered entrepreneurial aspirations are shaped by a constellation of contextual influences that unfold across the life course. Early-life socialisation, schooling and cultural norms shaped gendered entrepreneurial pathways (López-Delgado, Iglesias-Sánchez & Jambrino-Maldonado 2019; Sweida et al. 2022; Vracheva & Stoyneva 2020). In adulthood, institutional frameworks, regulatory norms and social recognition continue to differentiate men’s and women’s entrepreneurial aspirations (Ali, Jabeen & Burhan 2023; Wu & Li 2020).

Women often face compounded barriers due to limited access to funding and networks, particularly in contexts such as South Africa (Jali et al. 2023; Shambare 2013). These interlocking influences demonstrate that entrepreneurial intentions emerge from the dynamic interaction of multiple factors across gender.

Research gap

Entrepreneurial intention is inherently gendered, shaped by perceptual, institutional and social factors, and cannot be explained by single-cause models (Nikou et al. 2019). Despite global research, three critical gaps remain in the South African context

Firstly, there is a lack of nationally representative, intersectional research capturing how demographic, perceptual and contextual determinants across genders interact with these factors. Secondly, existing research is mainly limited to students or urban populations, which excludes broader population groups, leading to narrow sampling. Lastly, limited analytical approaches in prior research fail to account for multifactorial interactions among determinants.

This study addresses these gaps by applying an integrated theoretical framework to the 2019 GEM dataset, which is the latest publicly accessible wave at the time of writing, offering a context-specific, analytical and intersectional examination of the gendered determinants of entrepreneurial intentions in South Africa.

Research methods and design

The sections that follow outline the study’s design, sample, measurement scales, data analysis procedures, reliability and validity evaluations and ethical considerations. The methodological objective of this study is to examine whether and how demographic, perceptual and contextual factors differentially shape entrepreneurial intentions among men and women in South Africa. To achieve this objective, the study employs a gender-disaggregated analytical strategy using nationally representative survey data and multivariate modelling techniques appropriate for a binary outcome variable.

Design

The study employs a cross-sectional survey design, leaning on secondary data from the 2019 GEM Adult Population Survey (APS) for South Africa. A cross-sectional design is appropriate for this study because the primary aim is not causal inference but the identification and comparison of associative patterns in entrepreneurial intentions across gender at a specific point in time. This is consistent with prior GEM-based entrepreneurship research.

The APS represents the most recent publicly accessible nationally representative dataset on entrepreneurial attitudes and activities in the country. The use of a publicly accessible, nationally representative dataset enhances transparency, replicability and comparability with prior studies. Although more recent data exist internally within GEM consortium, the 2019 dataset is the latest publicly released wave and remains methodologically robust for examining relationships between variables. The 2019 APS remains analytically valid for examining structural and perceptual determinants of entrepreneurial intentions, which tend to be relatively stable over short time horizons.

Sample

The GEM consortium uses stratified random sampling to ensure that it accurately represents the national population in key demographics such as gender, age, location, race and income categories. Global Entrepreneurship Monitor’s sampling strategy aims for a minimum of 2000 respondents per country to achieve national representativeness and adequate statistical power for sub-group comparisons. The South African respondents consisted of 1475 males and 1516 females who provided complete responses on entrepreneurial motivation variables. The available sample is sufficiently large and statistically appropriate for the analytical goals of this study.

Measurement scales

The APS targets adults aged 18 and older, as this demographic includes potential entrepreneurs and individuals actively involved in entrepreneurial activities. The APS is designed to collect detailed information about the entrepreneurial activities, attitudes and aspirations of respondents using GEM standard protocols. The survey captures key variables such as business characteristics, motivations for starting a business, steps taken to establish and manage ventures and attitudes towards entrepreneurship. For detailed information on sampling methodology and data collection, refer to the GEM website (see www.gemconsortium.org). For this study, only some of the measured variables were used. The dependent variable, future entrepreneurial intentions, measures an individual’s intention to engage in entrepreneurial activities. Independent variables were age, household income, self-efficacy, fear of failure, knowledge of entrepreneurs, entrepreneurial prestige, perceived ease of starting a business, perceived opportunities, entrepreneurship as a career and media representation of entrepreneurship.

The items and scales used to collect data are summarised in Table 1.

Data analysis

A binary logistic regression model was used to analyse whether a selection of demographic, perceptual and contextual factors on entrepreneurial intentions had a significant effect on the odds of observing the ‘Yes’ category of entrepreneurial intentions. The appropriateness of this approach was based on a number of assumptions. Firstly, the dependent variable, future entrepreneurial intentions, was dichotomised, satisfying the model requirement. Secondly, observations are assumed to be independent, which is appropriate given the cross-sectional survey design. Thirdly, multicollinearity among predictors was assessed using variance inflation factors (VIF), with all values falling below the accepted threshold, indicating no significant multicollinearity concerns. While logistic regression does not assume normality of predictors, it assumes linearity in the logit for continuous variables. Given the categorical and ordinal nature of the variables used, this assumption is reasonably satisfied. These diagnostics support the appropriateness and robustness of the analytical approach.

The purpose of the test was to examine the relationship between the variables specified in Table 1. Separate models were estimated for males and females, using Intellectus Statistics software to allow for direct comparison of effect sizes and significance patterns, rather than imposing equality constraints across gender.

Following Hosmer, Lemeshow and Sturdivant (2013), model fit and explanatory power were assessed using model Chi-square (significance indicates improvement over the null model), McFadden’s R² (pseudo-R² measure of explanatory strength) and Wald Chi-square statistics for testing the unique significance of each predictor (p < 0.05).

Odds ratios (ORs) were interpreted for significant predictors only, using OR = 1 as the reference point. Deviations above or below 1 were expressed as percentage increases or decreases in the likelihood of entrepreneurial intentions.

Reliability and validity

Reliability in this study is supported through the standardised design and administration of the GEM APS, which follows rigorous cross-national data collection protocols. The use of consistent measurement instruments across respondents enhances the stability and comparability of responses. Furthermore, the large sample size and stratified sampling approach contribute to the robustness and reproducibility of the findings. Although single-item measures limit internal consistency assessment, their widespread use in GEM-based research supports their reliability for capturing broad perceptual constructs. Quality assurance was further addressed through diagnostic testing for multicollinearity using VIF (see Table 2). Together, these measures support the construct validity and analytical integrity of the models.

All predictors recorded VIF values below 2, well under the commonly accepted threshold of 5 (Menard 2009), confirming that multicollinearity was not a concern (see Table 2). The highest observed VIF was 1.79.

Ethical considerations

Ethical clearance to conduct this study was obtained from University of South Africa’s Graduate School of Business Leadership and Research Ethics Committee (Ref. No. [2022_ SBL_AC_001_SD]).

Results

The results are presented under the headings demographic data and then the regression data as it pertains to men and the women.

Demographic

Table 3 summarises the demographic profile of the respondents.

The respondents represented a diverse age range but were predominantly younger adults (18–34 years). The majority were not currently working (32.7%), while those fully employed (22.6%) or self-employed (10.2%) together roughly matched the proportion of unemployed individuals. The proportion of part-time workers was also relatively low (6.6%). Respondents were relatively well educated, with 40.3% having completed secondary school and almost 40% holding qualifications beyond high school (23.0% with post-school certificates or diplomas and 9.0% with university degrees). Household income was evenly distributed, likely as a result of the GEM consortium’s sampling strategy.

Males sub-group: Logistic regression model

The logistic regression model for males was statistically significant, χ²(38) = 141.36, p < 0.001, with McFadden’s R² = 0.13. This indicates moderate explanatory power, which is 13% of the variance in predicting future entrepreneurial intentions.

In Table 4, among the predictors in the model, only age, skills (self-efficacy) and know (knowledge of entrepreneurs) were significant predictors of entrepreneurial intentions:

• Compared to the reference group (18–24 years), men aged 45–54 years had 51% lower odds of entrepreneurial intentions (OR = 0.49, p = 0.018). Men aged 65–120 years showed a 79% reduction in odds (OR = 0.21, p = 0.013).
• Furthermore, men who strongly agreed they had the skills to start a business had 76% higher odds (OR = 1.76, p = 0.024) compared to those who strongly disagreed.
• Lastly, knowing 1 entrepreneur doubled the odds (OR = 2.97, p < 0.001), 2–4 entrepreneurs tripled the odds (OR = 3.39, p < 0.001) and ≥ 5 entrepreneurs increased odds eightfold (OR = 8.00, p < 0.001) compared to knowing none.

However, household income, fear of failure, perceived ease of starting a business, perceived opportunities, entrepreneurial prestige, media representation of entrepreneurship and entrepreneurship as a career showed no significant effects (p > 0.05).

Females sub-group: Logistic regression model

Table 5 shows that the binary logistic regression model (female respondents) predicting future entrepreneurial intentions was significant, χ²(38) = 144.50, p < 0.001. This means the predictors collectively had a significant influence on the odds of future entrepreneurial intentions). The model’s fit was acceptable (McFadden’s R² = 0.16).

Age, perceived skills, fear of failure and entrepreneurial knowledge were significant predictors of future entrepreneurial intention:

• Concerning age, compared to the 18–24 years age group, females aged 25–34 years (OR = 0.49, p = 0.008), 55–64 (OR = 0.36, p = 0.009) and 65–120 (OR = 0.27, p = 0.042) had significantly lower odds of future entrepreneurial intentions.
• In terms of perceived skills (self-efficacy), females who ‘strongly agreed’ had the necessary skills and had significantly higher odds of future entrepreneurial intentions (OR = 2.19, p = 0.004).
• As for fear of failure, females who ‘strongly agreed’ feared failure and had significantly lower odds of future entrepreneurial intentions (OR = 0.53, p = 0.016).
• Lastly, in terms of know (knowledge of entrepreneurs), knowing one entrepreneur increased the odds fourfold (OR = 4.12, p < 0.001) or 2–4 entrepreneurs more than three-and-a-half-fold (OR = 3.62, p < 0.001) known entrepreneurs significantly increased the odds of future entrepreneurial intentions compared to having none.

However, the variables household income, perceived social status, perceived ease of starting a business, perceived opportunities, perceived community support and media influence did not significantly predict future entrepreneurial intentions in this model.

In summary, Table 4 (men) and Table 5 (women) show that future entrepreneurial intentions of both men and women are influenced by age, self-efficacy and the number of entrepreneurs they know. However, women are negatively affected by fear of failure.

Discussion

Prior studies have emphasised the importance of promoting entrepreneurial activity among working-age adults as a means to empower them socio-economically and enhance economic activity (Khursheed 2022; Omeje et al. 2020; Prasannath et al. 2024). This study contributes to this discourse by examining gender-linked variations in the factors which influence entrepreneurial intentions among South Africans using nationally representative data. Both male and female logistic regression models were statistically significant. However, the explanatory power was slightly higher for the model for females (McFadden’s R² = 0.16) compared to the one for males (McFadden’s R² = 0.13). This demonstrates a moderate fit and gender differentiated influences.

In addressing the first research question, both subgroups showed that entrepreneurial intentions decreased with age, although at different stages. For males, the decrease began after 45 years, while for females, it started earlier (from 25 to 34 years). This pattern aligns with explanations rooted in life-course theory where older individuals often experience increased risk aversion, greater role commitments or entrenched career pathways that reduce motivation to pursue entrepreneurship (Paulsen et al. 2012). The earlier decline in women’s intentions may reflect gendered social roles including caregiving expectations, domestic responsibilities or societal pressures that reduce their risk-taking latitude (Adikaram & Razik 2023; Strawser et al. 2021). These findings reinforce gendered opportunity structures within the South African labour market where women frequently face constrained economic agency.

Consistent with Ali et al. (2023), household income was not a significant predictor of entrepreneurial intentions for either men or women. This contrasts with resource-based and human capital perspectives, which assume that higher income facilitates entrepreneurial entry by reducing risk and increasing investment capacity. In the South African context, however, income showed no meaningful effect. Although often associated with necessity-driven entrepreneurship, lower-income groups in this study did not demonstrate higher entrepreneurial intention.

With regard to the second research question, self-efficacy emerged as a significant predictor of entrepreneurial intentions for both genders but with a stronger effect for women than men. This finding reinforces SCT, which posits that perceived capability is a central driver of behavioural intention. The stronger effect observed among women suggests that self-efficacy may function as a compensatory mechanism in contexts where structural barriers are more pronounced.

The stronger effect observed among women contrasts with prior studies that have reported self-efficacy as either more influential factor for men (Ali et al. 2023), or no significant gender differences on the effects of self-efficacy on entrepreneurial intentions (Song & Lu 2024). This suggests that for South African women, confidence in their entrepreneurial competence may play a disproportionately enabling role.

The fear of failure significantly reduced entrepreneurial intentions for women but not for men. This suggests that this psychological constraint operates differently across gender. This effect may be explained by the interaction between internalised risk perceptions and structurally reinforced insecurity within the labour market, where women face greater penalties for failure. In the South African context, characterised by high economic precarity and limited safety nets, the perceived cost of failure may be disproportionately higher for women, thereby suppressing entrepreneurial entry. This aligns with GRT, which suggests that socially constructed gender norms shape differential risk perceptions and behavioural responses.

The finding is consistent with Ali et al. (2023) and aligns with the GRT, which argues that women internalise greater self-doubt and heightened risk sensitivity due to social sanctions, expectations of caution and lower tolerance for risk. However, this is in contrast to Wannamakok and Chang (2020) who found no gendered effect. These inconsistencies call for greater attention to context-specific psychological barriers in emerging market settings where structural inequalities may amplify fear among women.

The results also show that both the male and female respondents benefited from knowledge of entrepreneurs. This supports the view that networks and role models serve as critical sources of vicarious learning and opportunity awareness (Ratten, Ferreira & Fernandes 2016; Wasim et al. 2024). However, the gendered pattern implies different social network structures. This finding aligns with gendered social capital literature, which highlights women’s restricted access to diverse entrepreneurial networks.

Contrary to the predictions of the GEM model, the theory of planned behaviour (Ajzen 1991) and SCT, entrepreneurial prestige, perceived ease of starting a business, perceived opportunities, entrepreneurship as a career and media representation of entrepreneurship did not significantly predict entrepreneurial intentions. This suggests that these perceptions may be less relevant in environments where structural constraints, institutional failures and high unemployment overshadow abstract perceptions.

The results have important theoretical implications. Notably, they challenge the universality of entrepreneurship models whose constructs presume relatively stable institutional environments. It also suggests the need for contextualised intention models that account for inequality, informal labour markets and high necessity entrepreneurship, featuring characteristic of South Africa. In addition, they highlight that policy assumptions linking opportunity perception or media visibility to entrepreneurial behaviour may not hold in contexts marked by socio-economic precarity. The study also advances the integration of SCT, IT and GRT by demonstrating that gendered entrepreneurial intentions are not driven by isolated psychological or structural factors but by their combined, contextually contingent effects. Lastly, the findings refine existing gender-based entrepreneurship models by identifying fear of failure as a gender-specific constraint and self-efficacy as a disproportionately enabling factor for women, thereby highlighting the need for more context-sensitive and gender-differentiated theoretical frameworks. Collectively, these contributions extend the current understanding of entrepreneurial intentions by demonstrating that gender effects are neither uniform nor universal but are shaped by the interaction among individual cognition, institutional conditions and socio-cultural expectations in emerging economy contexts.

Practical implications

The evidence from this study indicates a need for interventions targeting both men and women in South Africa to mitigate challenges surrounding entrepreneurial activity. The decline in entrepreneurial intentions among men after the age of 45 highlights the importance of promoting entrepreneurship as a viable late-career option. This pattern may reflect underlying structural barriers that constrain entrepreneurial entry later in life. While youth entrepreneurship remains important, targeted efforts to upskill older adults, who may possess valuable experience but lack contemporary business competencies, are equally necessary.

The influence of role models in shaping male entrepreneurial success also warrants attention. Gender-inclusive policies should therefore be refined to address social barriers and ensure equitable access to entrepreneurial networks and role-model exposure for both men and women. Strengthening these pathways could help reduce gendered differences in entrepreneurial confidence and engagement.

The study further reveals the debilitative effects of fear of failure among South African women, underscoring the need for supportive initiatives that promote confidence and resilience. Importantly, women with greater entrepreneurial knowledge displayed higher intentions to start a business, suggesting that access to relevant information and skills training may counteract fear of failure. Enhancing knowledge-based interventions may therefore be an effective mechanism for increasing entrepreneurial intentions among women.

Given the complexity of entrepreneurship and the influence of economic strata on individuals’ behavioural intentions, research of this nature could shift its emphasis from gender differences to the needs of people across different economic levels. Rather than comparing only two groups – men and women – examining three income-based groups (low, middle, and high income) may offer more meaningful insights and sufficient variation to conduct statistical analyses with greater confidence. These results may provide clearer information assisting those in need most.

Conclusion

The study examined how demographic, perceptual and contextual factors shape gendered differences in entrepreneurial intentions among South African adults using GEM 2019 APS data. The findings show that gender matters in some, but not all, aspects of entrepreneurship. While many drivers operate similarly across genders, certain intentions and influencing factors differ. Self-efficacy and knowing entrepreneurs were important predictors for both men and women, but women’s intentions were more affected by self-belief and fear of failure, whereas men benefited more consistently from broader entrepreneurial networks. Gender appears only as a marginal influencer of entrepreneurial intention. Rather than replacing existing theories, this study suggests that they should be contextualised and extended to capture gender-specific mechanisms relevant to emerging economies such as South Africa.

The study had some notable limitations. Firstly, the cross-sectional design restricts causal inference, limiting the ability to determine the directionality of relationships between predictors and entrepreneurial intentions. Secondly, the reliance on self-reported data introduces potential response biases, including social desirability and perceptual bias, which may affect the accuracy of reported attitudes and intentions. Thirdly, the use of GEM variables, while standardised and widely validated, constrains the inclusion of deeper socio-cultural, institutional and psychological factors, potentially limiting the explanatory power of the model. This is reflected in the relatively modest pseudo-R² values, indicating that additional unobserved variables may play a significant role. Finally, while nationally representative, the dataset does not fully capture dynamic changes over time, suggesting the need for longitudinal research to better understand the evolution of entrepreneurial intentions across gender. Future studies should draw on data that incorporates other relevant and often overlooked factors, such as personality traits (e.g. locus of control, resilience), family background, context, institutions and prior business experience, whose exclusion may limit the predictive accuracy of existing models.

Acknowledgements

The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

Takawira M. Ndofirepi: Conceptualisation, Formal analysis, Investigation, Methodology, Project administration, Software, Writing – original draft. Renier Steyn: Conceptualisation, Formal analysis, Supervision, Writing – review & editing. All authors reviewed the article, contributed to the discussion of results, approved the final version for submission and publication and take responsibility for the integrity of its findings.

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

The data used for this study are publicly available from the Global Entrepreneurship Monitor website - https://www.gemconsortium.org/data.

The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency or that of the publisher. The authors are responsible for this article’s results, findings and content.

References