This study aims to reexamine the moderating role of human capital on the effect of extended financial inclusion (FI) for entrepreneurship, using data from the Global Entrepreneurship Monitor for a sample of 42 countries from 2006 to 2017.
This study distinguished between actual and perceived human capital. Actual human capital was measured through formal education while perceived human capital was captured by self-perceived capabilities for business start-ups. The moderating role of human capital was captured by the interaction terms between FI and human capital to investigate how the effects of FI on entrepreneurship vary with levels of human capital. The estimation used the panel-corrected standard error estimators and the two-step system generalized method of moments estimators.
Higher levels of formal education decrease the positive effect of extended FI on entrepreneurial activities. Individuals with high levels of self-capability do not leverage FI for entrepreneurial activities as much as those with lower levels of perceived capability. The results are robust to different estimation methods and different forms of actual human capital.
Both financial and human capital matter for new business formation worldwide. The findings suggest that FI policies must account for the decreasing effect in response to high levels of human capital. Future research should explore different measures of entrepreneurial performance, various types of entrepreneurship and entrepreneurship across gender groups to gain deeper insights into strategies for promoting entrepreneurship.
Education strategies should focus on specific types of education, such as entrepreneurship education with financial literacy, rather than traditional academic curriculum, to foster entrepreneurship knowledge, skills and creativity. Likewise, entrepreneurship support schemes should aim to nurture and share appropriate levels of self-efficacy, avoiding excessively high self-efficacy, which is deleterious to the benefits of FI for entrepreneurial activities.
This study offers novel evidence of the decreasing effects of FI on entrepreneurial activities in response to increased actual and perceived human capital.
1. Introduction
Entrepreneurship is a key driver of productive activities and innovation, contributing to wealth creation through new economic opportunities (Wong et al., 2005) and reduced poverty, especially among countries with low income and high inequality (Asongu and Odhiambo, 2023). Growth in entrepreneurial activities is influenced by various factors, including financial capital (Gregory, 2019; Demirgüç-Kunt and Singer, 2017), technology development and exchange (Nguyen Chau et al., 2024), business opportunities and policy support (Schwab, 2018) and human capital associated with the ability to identify and take on new business opportunities (Koudstaal et al., 2016; Hayward et al., 2006).
Among these factors, financial capital plays a critical role, particularly through financial inclusion (FI) ensuring access to affordable financial services for individuals and businesses (Barajas et al., 2020; Gregory, 2019). This has become an important policy goal for fostering entrepreneurship, especially in developing economies (Asongu and Odhiambo, 2023; Corrado and Corrado, 2017), where improved access to financial services such as credit, savings and insurance enables entrepreneurs to overcome resource constraints and capitalize on business opportunities. FI has been linked to new venture formation, job creation and social welfare across nations (Kim et al., 2018; Grohmann et al., 2018; Demirgüç-Kunt and Singer, 2017). This is the key rationale for efforts to improve the inclusiveness of financial systems among developing countries (Corrado and Corrado, 2017).
Despite a strong theoretical foundation, the positive effect of FI on entrepreneurship is not universally consistent. Empirical studies report mixed effects of FI on nascent entrepreneurship across regions and countries, including developed countries in Europe and the USA (Grilo and Thurik, 2005), developing countries in Africa (Ajide, 2020) and China (Cao and Zhang, 2022; Beck et al., 2015). This inconsistency creates challenges in effectively allocating financial resources to support entrepreneurial growth. A possible reason for these varied effects is the role of human capital as a moderating factor. Although FI provides necessary resources, the way nascent entrepreneurs leverage them depends on their human capital.
There is also a rich literature discussing the effect of human capital on entrepreneurship. The literature distinguishes between actual human capital – measurable attributes such as formal education and technical skills – and perceived human capital or an individual’s self-assessed capabilities (Dutta and Sobel, 2018; Liao and Welsch, 2008; Reynolds et al., 2005; Davidsson and Honig, 2003). The human capital theory (Becker, 1964) and hubris theory (Hayward et al., 2006) both suggest that while human capital is important in driving economic growth, higher levels of human capital are not always associated with increased entrepreneurial activities.
Previous research has extensively examined the relevance of both FI and human capital across different country contexts and in relation to various outcomes. For example, the cross-country study by Du et al. (2022) for 18 emerging countries demonstrates that environmental quality improves with increased FI but deteriorates with human capital. Similarly, Saydaliyev et al. (2022) demonstrate that for remittance-receiving countries, FI and human capital both promote economic growth. Several studies have also explored the relationship between FI and human capital. For instance, Thathsarani et al. (2021) conclude that while FI has a significant short-term impact on economic growth, it negatively affects the development of human capital in South Asian countries. In contrast, Huang et al. (2023) find that higher levels of FI, combined with reductions in income inequality, tended to enhance human capital in 36 sub-Saharan countries between 2004 and 2019.
However, the interaction between FI and human capital remains underexplored, leaving a critical gap in understanding how these factors interact to shape entrepreneurial outcomes. A notable exception is the work of Boachie and Adu-Darko (2024), which explores how human capital development interacts with FI to drive economic growth. They find evidence supporting the positive moderating role of human capital, as its fundamental traits and acquired skills enhance both FI and economic growth. Despite insightful findings, the study does not account for different types of human capital attributes, focusing only on human capital stock. Moreover, their analysis is focused on economic growth rather than entrepreneurial outcomes, which could represent a specific conduct contributing to broader economic growth.
This study addresses this knowledge gap by examining the moderating role of human capital in the FI –entrepreneurship nexus. We propose that although FI generally supports entrepreneurship by reducing financial constraints, its effects may diminish as the levels of actual and perceived human capital increase. Specifically, individuals with higher levels of formal education (actual human capital) may face decreased incentives to pursue entrepreneurial activities, as education increases either their expected returns in entrepreneurial pursuits or risk aversion. Likewise, individuals with higher levels of perceived human capital may overestimate their self-capability and, as a result, underuse financial services designed to support entrepreneurship.
To investigate these relationships, this study uses data from the Global Entrepreneurship Monitor (GEM) for a sample of 42 countries spanning 2006–2017. The GEM data is available for measurements of nascent entrepreneurial activities, innovative entrepreneurial activities and perceived human capital through self-perceived capabilities. We gather data on actual human capital, such as formal education and the human development index and several controls, from World Development Indicators. FI is captured through a composite FI index following Sarma’s (2012) approach using the data from the International Monetary Fund’s Financial Access Surveys. Our econometric models for entrepreneurial activities incorporate interaction terms between FI and different forms of human capital to investigate how the effects of FI on entrepreneurship vary with levels of human capital. We apply the panel-corrected standard error (PCSE) estimator to control for heteroskedasticity. The study also uses the two-step system generalized method of moments (GMM) estimators to confirm the robustness of our results.
Our paper contributes to the existing literature on entrepreneurship and FI in several ways. First, we confirm the widely held belief that FI promotes entrepreneurship and entrepreneurial innovation among nascent entrepreneurs. Second, we extend the literature by highlighting that the effects of extended FI on entrepreneurship decrease among those with higher levels of formal education. In addition, we build on the hubris theory to highlight that a higher level of self-perceived capability decreases the effect of extended FI for entrepreneurship and entrepreneurial innovation. The findings will help inform government policies for monitoring the effects of FI on entrepreneurial outcomes.
The remainder of this paper is structured as follows. Section 2 reviews the literature on FI, human capital and entrepreneurship and develops the theoretical framework for the moderating role of human capital. Section 3 presents the data and empirical strategy, whereas Section 4 discusses the main results and robustness checks. Finally, Section 5 concludes the study.
2. Literature review and hypothesis construction
Entrepreneurship is crucial for economic growth, driving innovation and new business opportunities (King and Levine, 1993). It involves the ability of individuals to recognize and act on business opportunities, which requires a certain level of financial and human resources. Improved financial access assists in mobilizing and allocating necessary resources to new entrepreneurs, allowing them to establish new businesses (King and Levine, 1993). Human capital matters as it provides the skills and knowledge required to perform entrepreneurial tasks (Davidsson and Honig, 2003). The following sections elaborate on the role of financial capital and human capital in fostering entrepreneurship.
2.1 Financial inclusion and entrepreneurship
Financial capital, which includes the availability and accessibility of financial resources, is essential to nurturing and promoting entrepreneurship in both developing and developed countries (Barajas et al., 2020; Demirgüç-Kunt and Singer, 2017). FI supports entrepreneurship by providing nascent entrepreneurs with necessary financial resources (Dutta and Sobel, 2018; Bruhn and Love, 2014), encouraging well-planned business strategies and fostering efficient use of resources (Jiménez et al., 2015). Several cross-country studies document the positive effects of FI across continents and countries, including Africa (Boachie and Adu-Darko, 2024; Ajide, 2020), developing Asian countries (Chowdhury and Chowdhury, 2024; Charfeddine and Zaouali, 2022) and Europe (Oyewole et al., 2024). However, the relationship between FI and entrepreneurship remains inconclusive. Although many studies support a positive link (Barajas et al., 2020), others suggest ambiguity in this relationship (Beck et al., 2015; Grilo and Thurik, 2005).
The mixed findings on the role of FI on entrepreneurial outcomes can be attributed to differences in country cases, methodologies used and the absence of contextual factors governing the relationship. Most studies focus on single dimensions of FI, such as financial accessibility, financial availability or the usage of financial services (Angelucci et al., 2015; Bruhn and Love, 2014). Although each dimension of FI is essential, a comprehensive metric considering all three dimensions can provide a more holistic and accurate assessment of the financial system’s ability to support entrepreneurs and businesses (Cao and Zhang, 2022).
The first aim of this study is to reaffirm the positive role of FI toward entrepreneurship, using a composite index of all three dimensions, including accessibility, availability and usage. The study proposes the following hypothesis:
Financial inclusion will promote entrepreneurial activities.
2.2 Human capital and entrepreneurship
Becker’s (1964) human capital theory emphasizes the importance of education and training in fostering entrepreneurial growth and economic success. In entrepreneurship, human capital encompasses skills, knowledge and experience needed to identify and exploit business opportunities and to foster entrepreneurial alertness and endeavors (Dutta and Sobel, 2018; Brush et al., 2017; Tang and Murphy, 2012). Strong human capital supports entrepreneurs in creating comprehensive business plans, managing risks and strategically positioning long-term ventures (Baum et al., 2001). Furthermore, robust human capital enhances entrepreneurs’ ability to secure the necessary resources, particularly valuable in the early stages of a venture (Brush et al., 2001). Recent research distinguishes between actual and perceived human capital, expanding Becker’s original concept.
Actual human capital refers to measurable attributes an individual possesses, such as formal education, work experience and specific technical skills. Formal education, in particular, is frequently explored in relation to entrepreneurship (Ahn and Winters, 2023; Debarliev et al., 2022). It equips individuals with the skills and knowledge necessary for entrepreneurial success, fostering a growth mindset and enhancing entrepreneurial intentions and performance. Several cross-country and country-level analyses demonstrate that countries with higher levels of education are associated with greater growth in entrepreneurship ratio (Ahn and Winters, 2023; Debarliev et al., 2022; Unger et al., 2011; Davidsson and Honig, 2003). The analysis for European countries by Millán et al. (2014) further suggests that entrepreneurship education instills creative and innovative abilities essential for entrepreneurship and innovation.
Perceived human capital involves individuals’ self-assessment of their skills and capabilities (Gielnik et al., 2020; Brush et al., 2017). It is influenced by subjective perceptions and confidence in one’s abilities, which can shape entrepreneurial motivation and actions. Those who believe they have the necessary skills and knowledge to start a business are more inclined to seek and use financial resources for entrepreneurship (Hartog et al., 2010) and thus provide confidence to undertake the risk of entrepreneurship (Maczulskij and Viinikainen, 2023; Liao and Welsch, 2008; Reynolds et al., 2005). Empirical evidence shows that perceived entrepreneurial ability motivates individuals to take concrete actions, transitioning from the idea phase to entrepreneurship success (Bachmann et al., 2021; Gielnik et al., 2020; Townsend et al., 2010; McGee et al., 2009), especially among women (Brush et al., 2017). Recent cross-country analyses using GEM data by Yusuf et al. (2024) and Ali et al. (2023) further confirm that self-perceived capabilities are associated with greater growth in business start-ups across a range of countries. This study seeks to reaffirm the significant contribution of actual and perceived human capital to entrepreneurship. It presents the following hypotheses:
Formal education, as a measure of actual human capital, will promote entrepreneurial activities.
Self-capability for start-ups, a measure of perceived human capital, will promote entrepreneurial activities.
2.3. A further look into human capital
Although FI and human capital individually drive entrepreneurial activities, their interplay remains underexplored. This section draws upon the existing literature to examine how actual and perceived human capital might moderate the effect of FI on entrepreneurship.
2.3.1 The moderating role of actual human capital.
Formal education, a notable measure of actual human capital, is often seen as enhancing the effects of FI on entrepreneurial growth by improving financial literacy, skills, knowledge and attitudes needed to make informed decisions about financial matters (Millán et al., 2014). However, higher education levels may also dampen the benefits of extended FI for entrepreneurship.
There are several reasons for this. First, higher formal education implies greater expected returns on human capital investment (Becker, 1964). Facing the same level of opportunity, including those offered by financial services, those populations with higher levels of formal education may be less willing to pursue an entrepreneurial aspiration, as the perceived return on business opportunities may seem less attractive than employment opportunities (Malchow-Møller et al., 2010). Second, higher levels of formal education can increase risk aversion by raising risk awareness for individuals (Jung, 2015), thereby reducing the chance of pursuing an entrepreneurial opportunity (Iyigun and Owen, 1998). Third, formal education might hinder the creative thinking necessary for entrepreneurship and innovation. As young people progress through school, individuals may become more focused on finding “correct answers” rather than exploring innovative ideas. Cross-country evidence by Jiménez et al. (2015) shows that entrepreneurship rates are significantly lower among countries with high education levels, indicating that efforts to promote entrepreneurship would need to consider the role of education.
In brief, extensive formal education and its potential consequences, including an increased risk aversion, higher expected return on investment and reduced creative thinking, can mitigate the positive impact of FI on entrepreneurial activities. This leads to the following hypothesis:
High levels of formal education will weaken the effect of financial inclusion on entrepreneurial activities.
2.3.2 The moderating role of perceived human capital.
FI undoubtedly eases financial constraints for entrepreneurial growth. However, its effect might depend on the self-perceived human capital of the populations the financial system is targeting. We develop from the hubris entrepreneurship theory to suggest that high levels of self-capability or overconfidence may weaken the positive effects of FI on entrepreneurial growth.
The hubris entrepreneurship theory proposes that self-perceived capabilities add value to early-stage entrepreneurship (Hayward et al., 2006). This is because high levels of self-capability can result in a cognitive bias, leading entrepreneurs to overestimate their chances of success and underestimate both the resources needed and potential risks involved in starting a new venture (Singh, 2020; Gielnik et al., 2020; Brush et al., 2017; Robinson and Marino, 2015). Therefore, overconfident entrepreneurs, driven by energy and commitment, often start businesses with limited resources (Robinson and Marino, 2015; McGee et al., 2009; Koellinger et al., 2007; Hayward et al., 2006), making them less reliant on financial resources than those with lower self-perceived capabilities (Tsai et al., 2016). Similarly, when tracking entrepreneurial growth, the number of start-up businesses often surges during the early stages when financial resources are scarce. However, this growth tapers off as resources become more plentiful over time. The recent cross-country analysis using GEM data by Brush et al. (2017) demonstrates that countries and populations with high levels of self-perceived capabilities often exhibit higher levels of new start-ups. However, this relationship cannot be found in established firms.
In light of these arguments, we propose the following hypothesis:
Excessive perceived self-capability for start-ups will mitigate the effect of financial inclusion on entrepreneurial activities.
3. Data, empirical strategy and model specification
3.1 Data description
We collected and collated country-level data from multiple secondary sources. First, we used the GEM database to capture country-level data on entrepreneurship; perceived human capital; and perceived business opportunities and risks by potential entrepreneurs. These GEM data have been used by several recent cross-country studies focusing on self-perceived capabilities and entrepreneurship, including Yusuf et al. (2024), Sima (2023), Ali et al. (2023), Rietveld and Patel (2023), Brush et al. (2017), among others.
We captured nascent entrepreneurship and entrepreneurial innovations using two indicators from the Adult Population Surveys within the GEM data, which are consistent with our theoretical arguments in Section 2, and consistent with cross-country entrepreneurship analyses presented above. These include total early-stage entrepreneurial activity (TEA) rate, which measures the percentage of the 18–64 population who self-reported as a nascent entrepreneur or new business owner-manager and innovation rate among new entrepreneurs (INNO), which measures the percentage of individuals offering products or services which are new to at least some customers. Perceived human capital was gauged by the proportion of the population aged 18–64 who believe they have the skills and knowledge to start a business (Yusuf et al., 2024; Ali et al., 2023; Brush et al., 2017; Wong et al., 2005). Following Yusuf et al. (2024) and Ali et al. (2023), we controlled for the effect of the fear of business failure and perceived business opportunities, sourced from GEM data.
Second, we used the Financial Access Survey from the IMF for constructing the FI index (FI). We followed Sarma (2012) to construct a composite FI index encompassing accessibility, availability and actual usage of financial systems. These three dimensions of FI are necessary for assessing the condition of FI in each economy and tracking the success of policy measures aimed at promoting FI. This comprehensive measure has been increasingly adopted by several country-level and cross-country studies (Tram et al., 2023; Saha and Qin, 2023; Lenka, 2022; Cao and Zhang, 2022).
Third, we used the World Development Indicators to obtain country-level data on actual human capital and control variables for the analysis. Following recent studies into entrepreneurship, actual human capital was measured through formal education (Ahn and Winters, 2023; Debarliev et al., 2022). We controlled for country-level GDP per capita as a source of physical capital for entrepreneurial growth (Yusuf et al., 2024; Sima, 2023). Finally, we extended existing entrepreneurship literature by considering country-level trade openness as a source of technological innovation for entrepreneurial growth. Based on the latest data available, our final data consisted of an unbalanced sample covering 42 countries between 2006 and 2017. Table 1 presents a summary of variable definitions and sources.
Variable definitions and measurement
| Abbreviation | Variable | Definition | Source |
|---|---|---|---|
| TEA | New entrepreneurs (%) | Percentage of 18–64 population who are nascent entrepreneurs or managing a new business | GEM |
| INNO | Innovation rate (%) | Percentage of TEA offering new products/services to some customers | GEM |
| FI | Financial inclusion | Financial inclusion index | Author’s calculation from the financial access surveys (IMF) |
| H_CAP | Formal education (%) | Percentage of tertiary enrollments to the total population for each country | WDI |
| P_CAP | Perceived human capital (%) | Percentage of 18–64 population confident in their skills and knowledge to start a business | GEM |
| HDI | Human development index | Human Development Index | WDI |
| GDP | GDP per capita ($US) | Per capita GDP | WDI |
| Trade | Trade openness (%) | Sum of exports and imports as a percentage of GDP | WDI |
| FEAR_FAIL | Fear of failure (%) | Percentage of 18–64 population who indicate that fear of failure would prevent them from setting up a business | GEM |
| P_OP | Perceived opportunities (%) | Percentage of 18–64 population seeing good opportunities to start a firm locally | GEM |
3.2 Model specifications
Our baseline models for explaining country-level entrepreneurship (TEA, INNO) as a function of FI, human capital and other factors take the following forms:
where TEA and INNO are outcome variables representing the percentage of the population engaging in new businesses and new products/services, respectively. Xj,i,t-1 is a set of control variables influencing entrepreneurial growth, including GDP per capita as physical capital financing entrepreneurship (Yusuf et al., 2024; Sima, 2023; Gregory, 2019), Trade (trade openness) as a source of technology transfer and exchange (Nguyen Chau et al., 2024), P_OP (perceived business opportunities) to capture supporting policies and FEAR_FAIL (the fear of failure) to capture potential risks associated with business startups perceived by potential entrepreneurs (Yusuf et al., 2024; Ali et al., 2023). Following Rauch and Rijsdijk (2013) and Dutta and Sobel (2018), our independent variables were lagged to capture their delayed effects on entrepreneurial outcomes.
To examine how the effects on entrepreneurship of FI might differ in response to actual and perceived human capital, we extended our baseline models by introducing interactions between FI and H_CAP (actual human capital) and P_CAP (perceived human capital) stepwise using equations (3)–(6). Consistency in the estimates between models with and without interactions is important for interpreting our variables of interest. If the interactions between FI and H_CAP (Fi,t - 1*H_CAPi,t - 1) are statistically significant, we can conclude that the effects of FI on entrepreneurial outcomes (TEA or INNO) depend on the levels of education of the population. A similar interpretation applies to the interactions between FI and P_CAP. Furthermore, we also checked the robustness of these interactions by estimating a similar model using interactions between FI and the human development index (HDI):
3.3 Estimation techniques
This study used the PCSE estimators to estimate the parameters of equations (1)–(6). The PCSE estimators have been considered to provide the most reliable results for complex panel data where the data exhibit significant heterogeneity over time and across units (Bailey and Katz, 2011). This advantage from PCSE is particularly relevant for our panel because there are significant differences in unobserved factors potentially associated with growth in entrepreneurial activities across countries and across times. This PCSE method has been widely used by several researchers, including, but not limited to, Liu and Ma (2021) and Brodzicki et al. (2018).
Concerning alternative estimation methods, both the PCSE and GMM estimators allow autocorrelation within the panel, meaning that the growth of entrepreneurial activities of one country can be correlated over time. However, the GMM estimators generally fail to address within-country heterogeneity. This implies that if countries experienced significant changes in entrepreneurial growth and its driving factors, GEM estimators cannot predict entrepreneurial outcomes as effectively as their PCSE counterparts do. Therefore, the PCSE was our preferred method also this method might not be ideal if entrepreneurship within a country has its momentum over time. We also reported in Section 4 the GMM estimates using equations (1)–(6) as a robustness check.
4. Result discussions
4.1 Results for total early-stage entrepreneurial activity
Table 2 reports the estimation results of equations (1), (3) and 4 using TEA as the dependent variable. The estimates for the control variables are largely consistent in terms of sign, magnitude and statistical significance across models with and without interactions between FI and H_CAP and P_CAP. Trade openness is positively associated with greater entrepreneurial activities in all models, suggesting that a country’s openness to international trade is likely to foster domestic entrepreneurial growth.
PCSE regression results with total early-stage entrepreneurial activity as the dependent variable
| Model specification (1) | Model specification (2) | Model specification (3) | Model specification (4) | |
|---|---|---|---|---|
| L.FI | 0.438*** (0.077) | 1.051*** (0.083) | 0.804*** (0.128) | 1.153*** (0.070) |
| L.GDP | −0.248*** (0.028) | −0.398*** (0.027) | −0.254*** (0.024) | −0.574*** (0.035) |
| L.P_OP | 0.564*** (0.034) | 0.362*** (0.034) | 0.274*** (0.033) | 0.395*** (0.028) |
| L.FEAR_FAIL | 0.229*** (0.055) | −0.223*** (0.051) | −0.249*** (0.058) | −0.353*** (0.048) |
| L.Trade | 0.074 (0.140) | 0.302*** (0.115) | 0.331*** (0.123) | 0.252** (0.108) |
| L.H_CAP | 1.486*** (0.083) | |||
| L.FI#L.H_CAP | −0.279*** (0.021) | |||
| L.P_CAP | 1.090*** (0.092) | |||
| L.FI#L.P_CAP | −0.183*** (0.031) | |||
| L.HDI | 9.246*** (0.579) | |||
| L.FI#L.HDI | −1.369*** (0.092) | |||
| Number of observations | 291 | 256 | 291 | 291 |
| Number of groups | 41 | 37 | 41 | 41 |
| χ² | 13,285 | 27,637.8 | 21,169.6 | 26,564.1 |
***p < 0.01; **p < 0.05; *p <0 0.1. Standard errors in parentheses. The table also reports the χ2 of the modified Wald test for group-wise heteroscedasticity with the null hypothesis of homoscedasticity
The consistent negative estimates for GDP per capita suggest that countries with higher GDP per capita tend to have lower rates of nascent entrepreneurship. This seemingly counterintuitive finding can be explained by factors like high market competition, opportunity cost for self-employment and business entry costs (Audretsch et al., 2012). Recent research corroborates that developed countries typically exhibit low self-employment rates. A general decline in self-employment rates has been observed across 21 OECD European countries, particularly in Northern and Western Europe (Saridakis et al., 2019). These findings align with prior research showing that developed countries such as Germany, France, the USA, Denmark, Sweden, Norway, Netherlands and Finland have low self-employment rates below 10% (Piotr and Rekowski, 2009). Conversely, perceived business opportunities (P_OP) consistently show positive impacts (0.274–0.564) across all models. With GDP and Trade controlled for, P_OP may capture the effect of policy rewards aimed at generating a dynamic business environment within countries. Thus, our results for P_OP align with existing literature emphasizing the role of social capital in entrepreneurship growth (Audretsch et al., 2012).
The positive estimates for FI in Columns 1–4 suggest a complementary effect of FI on the TEA rate, confirming H1. This finding supports related literature contending that improved financial services release financial constraints, enabling potential entrepreneurs to start businesses (Gregory, 2019; King and Levine, 1993). Moreover, increased FI, through improved financial and banking services, reduces transaction costs and information asymmetry, resulting in more incentives to set up a new business (Charfeddine and Zaouali, 2022).
Both actual human capital (H_CAP) and perceived human capital (P_CAP) exhibit positive effects on TEA across the models, confirming H2a and H2b. Formal education nurtures entrepreneurial intentions and attitudes, enhances entrepreneurial skills and knowledge to manage ventures (Unger et al., 2011) and contributes to entrepreneurship performance. Perceived capabilities (P_CAP) are positively associated with greater TEA, confirming existing literature emphasizing the role of self-confidence in business start-ups and entrepreneurship decisions (Maczulskij and Viinikainen, 2023; McGee et al., 2009). Our international panel evidence reinforces the argument that higher levels of human capital create a more dynamic entrepreneurial environment (Brush et al., 2017; Brush et al., 2001).
However, the negative interaction between FI and H_CAP (−0.279) confirms H3. This result is consistent when we introduced interactions between FI and the human capital index (HDI) in Column 4. While both FI and actual formal education positively impact entrepreneurial growth and business startups, the effect of extended FI decreases among better-educated populations. Higher education is associated with increased risk aversion, which can act as a barrier to entrepreneurial intentions despite improved FI. Recent studies suggest that while higher education enhances entrepreneurial knowledge and competencies, it often coincides with a declining intention toward entrepreneurship due to a preference for more secure career paths (Setyanti, 2021; Iyigun and Owen, 1998). Our cross-country evidence contradicts Millán et al. (2014), who argue that better-educated people make better use of financial services for business opportunities.
Similarly, the negative interaction between FI and P_CAP (−0.183) supports H4. The effect of extended FI on entrepreneurial growth is lower among those with high levels of self-perceived capability, consistent with the hubris entrepreneurship theory. Potential entrepreneurs, driven by energy and commitment, often start businesses with limited resources (Robinson and Marino, 2015; Koellinger et al., 2007; Hayward et al., 2006), making them less reliant on financial resources than those with lower self-perceived capabilities (Tsai et al., 2016). Therefore, we expect to see lower effects when FI is improved in the populations with high levels of perceived capabilities. In addition, when observing populations with high levels of self-perceived capabilities, the number of start-up businesses often surges during the early stages when financial resources are scarce. However, this growth tapers off as resources become more plentiful over time.
4.2 Results for innovative total early-stage entrepreneurial activity
Next, we examine whether human capital reduces the positive effect of extended FI for innovative entrepreneurship (INNO) by estimating equations (2), (5) and (6). This variable captures whether new business start-ups involve new products and services. The results are reported in Table 3. Most control variables, including GDP, P_OP, yield consistent results with those presented in Table 2 where TEA was the dependent variable. The positive estimates for FI (FI), actual human capital (H_CAP) and perceived human capital (P_CAP) remain statistically significant, supporting H1, H2a and H2b. These findings confirm that FI, formal education and perceived capabilities are drivers of entrepreneurial growth, including innovative entrepreneurial activities.
PCSE regression results with innovative total early-stage entrepreneurial activity as the dependent variable
| Model specification (1) | Model specification (2) | Model specification (3) | Model specification (4) | |
|---|---|---|---|---|
| L.FI | 0.230*** (0.046) | 0.761*** (0.059) | 0.727*** (0.088) | 0.751*** (0.070) |
| L.GDP | 0.023 (0.018) | −0.064*** (0.018) | −0.026 (0.017) | −0.013 (0.037) |
| L.P_OP | 0.262*** (0.021) | 0.104*** (0.023) | 0.119*** (0.027) | 0.110*** (0.025) |
| L.FEAR_FAIL | 0.463*** (0.031) | 0.173*** (0.034) | 0.188*** (0.035) | 0.177*** (0.040) |
| L.Trade | 0.051 (0.096) | 0.036 (0.060) | 0.023 (0.053) | 0.043 (0.056) |
| L.H_CAP | 0.683*** (0.081) | |||
| L.FI#L.H_CAP | −0.149*** (0.014) | |||
| L.P_CAP | 0.522*** (0.058) | |||
| L.FI#L.P_CAP | −0.132*** (0.024) | |||
| L.HDI | 2.586*** (0.573) | |||
| L.FI#L.HDI | −0.690*** (0.074) | |||
| Number of observations | 239 | 216 | 239 | 239 |
| Number of groups | 40 | 36 | 40 | 40 |
| χ² | 102,500 | 136,540 | 122,014 | 103,054 |
***p < 0.01; **p < 0.05; *p < 0.1.
Standard errors in parentheses. The table also reports the χ2 and the p-value of a modified Wald test for groupwise heteroscedasticity with the null hypothesis of homoscedasticity
The interaction term between H_CAP and FI in Column (2) is statistically significant and negative (−0.149), suggesting that the positive effect of FI on entrepreneurial innovations decreases with higher levels of formal education. This finding still holds when H_CAP is replaced by the human development index (HDI) in Column (4). This result again confirms H3 stating that higher levels of formal education would decrease the (positive) effect of extended FI on entrepreneurial growth and innovations, aligning well with the earlier finding for TEA. This reinforces the view that extended FI for entrepreneurial growth and innovations would achieve less in populations with high levels of formal education, aligning with Setyanti (2021) and Wang et al. (2021), who demonstrate that formal education, often linked to increased risk aversion and reduced entrepreneurial intention, dampens the effect of financial resources on entrepreneurial activity.
Similarly, the estimate for the interaction between FI and P_CAP (self-perceived capabilities) is statistically significant and negative (−0.132), confirming our H4. The hubris theory is relevant to this decreasing effect of FI for entrepreneurial innovations. Taking it altogether, potential entrepreneurs, driven by self-confidence and commitment, often make business decisions, including the introduction of new products and services, even when the resources are limited (Robinson and Marino, 2015; Koellinger et al., 2007; Hayward et al., 2006). This growth tapers off as resources become more plentiful over time.
4.3 Robustness tests
We conducted a robustness test on the moderating effect of formal education (H_CAP) and self-perceived capabilities (P_CAP) using the GMM estimators. Specifically, we estimated GMM models on both TEA and INNO as the dependent variables following equations (1)–(6), and compared the estimates for the interaction terms between FI and H_CAP and P_CAP with those presented in Tables 2 and 3. The GMM models were estimated using a two-step system GMM procedure. The two-period lag of each dependent variable is included to account for the dynamics of entrepreneurial activities as the Arellano–Bond autocorrelation tests suggest no autocorrelation at the second order. Our GMM models satisfy the null hypothesis of no overidentification using the Sargan and Hansen-J tests. The results are presented in Table 4.
Regression results using two-step system GMM estimators
| TEA as the dependent variable | INNO as the dependent variable | |||||||
|---|---|---|---|---|---|---|---|---|
| Model specification (1) | Model specification (2) | Model specification (3) | Model specification (4) | Model specification (5) | Model specification (6) | Model specification (7) | Model specification (8) | |
| L2.Dependent variable | 0.326** (0.145) | 0.328*** (0.057) | 0.240*** (0.075) | 0.232*** (0.075) | 0.585*** (0.163) | 0.366*** (0.047) | 0.365*** (0.059) | 0.411*** (0.056) |
| L.FI | 0.581** (0.279) | 0.817*** (0.190) | 1.274*** (0.352) | 1.101*** (0.219) | 0.349 (0.242) | 0.928*** (0.080) | 0.682*** (0.146) | 0.748*** (0.067) |
| L.GDP | −0.265*** (0.094) | −0.310*** (0.058) | −0.364*** (0.068) | −0.459*** (0.132) | −0.078 (0.058) | −0.142*** (0.049) | −0.100*** (0.029) | −0.164*** (0.056) |
| L.P_OP | 0.487*** (0.122) | 0.404*** (0.062) | 0.449*** (0.063) | 0.403*** (0.055) | 0.144** (0.060) | 0.073*** (0.015) | 0.063 (0.044) | 0.006 (0.032) |
| L.FEAR_FAIL | −0.006 (0.306) | −0.179 (0.137) | −0.295* (0.157) | −0.277* (0.138) | 0.109 (0.088) | −0.046 (0.052) | 0.115 (0.076) | 0.055 (0.054) |
| L.Trade | 0.297 (0.209) | 0.344*** (0.101) | 0.388*** (0.105) | 0.365*** (0.099) | −0.111 (0.203) | −0.031 (0.056) | −0.076 (0.053) | −0.081** (0.031) |
| L.H_CAP | 0.448 (0.313) | 0.494*** (0.082) | ||||||
| L.FI#L.H_CAP | −0.085* (0.042) | −0.129*** (0.013) | ||||||
| L.P_CAP | 0.505*** (0.179) | 0.508*** (0.065) | ||||||
| L.FI#L.P_CAP | −0.157** (0.076) | −0.128*** (0.031) | ||||||
| L.HDI | 4.600* (2.293) | 3.483*** (0.885) | ||||||
| L.FI#L.HDI | −0.754*** (0.258) | −0.682*** (0.128) | ||||||
| N | 291 | 256 | 291 | 291 | 239 | 216 | 239 | 239 |
| AR(2) statistics | 1.177 | 1.359 | 1.405 | 1.410 | 1.273 | 1.559 | 0.764 | 0.831 |
| AR(2) p-value | 0.239 | 0.174 | 0.160 | 0.158 | 0.203 | 0.119 | 0.445 | 0.406 |
| Sargan statistics | 17.663 | 19.423 | 17.504 | 18.824 | 27.852 | 31.217 | 27.690 | 25.173 |
| Sargan’s p-value | 0.610 | 0.430 | 0.556 | 0.468 | 0.315 | 0.355 | 0.535 | 0.669 |
| Hansen-J statistics | 17.145 | 14.436 | 15.487 | 16.224 | 26.950 | 25.020 | 25.185 | 23.492 |
| Hansen’s p-value | 0.644 | 0.758 | 0.691 | 0.642 | 0.358 | 0.677 | 0.669 | 0.754 |
***p < 0.01; **p < 0.05; *p <0 0.1. Standard errors in parentheses. AR(2) tests suggest that autocorrelation is not a major concern when using the second lag of the dependent variable. All the p-values for Sargan and Hansen-J tests suggest that overidentification is not an issue in our GMM approach
Our GMM estimates in Table 4 for most control variables (GDP, P_OP) are consistent with those presented in Tables 2 and 3 in terms of statistical significance, sign and magnitude. The estimates for FI (FI), and two variables representing actual and perceived human capital (H_CAP, P_CAP) are also consistent, suggesting that these are drivers of entrepreneurial growth and innovations. The interaction terms between FI (FI) and formal education (H_CAP) and self-perceived capabilities (P_CAP) are statistically significant and negative, confirming the results from the PCSE estimates. Our findings for the nexus between FI, human capital and entrepreneurial growth are consistently robust, supporting the hypotheses developed from the existing literature.
5. Conclusions
This study explores how human capital influences the relationship between FI and entrepreneurship, using GEM data for a sample of 42 economies from 2006 to 2017. The study offers several unique features. First, FI is captured by a composite index encompassing accessibility, availability and usage of financial services. Second, the study examines both actual human capital, measured by formal education, and perceived human capital, measured by perceived self-capability for start-ups. Third, the study applies diverse theoretical arguments, from economics to risk aversion and entrepreneurship, to justify the human capital’s role in the FI–entrepreneurship nexus. Our robust results provide new evidence of this moderating effect.
Findings show that a more inclusive financial system enables new venture creation. However, the extent of this positive effect depends on the level of human capital. Interestingly, countries with higher human capital exhibit less pronounced effects than those with lower human capital. This holds for both actual human capital – formal education and the human development index – and perceived human capital – self-perceived capabilities of potential entrepreneurs. The result is consistent across entrepreneurial outcomes using the TEA and innovative TEA rates. These findings support the risk aversion argument, which considers that higher formal education is associated with greater risk aversion (Jung, 2015) and higher expected returns on human capital investment (Becker, 1964). The finding for the moderating role of perceived capabilities further lends credence to the hubris theory of entrepreneurship, which associates excessive perception of self-capability with overconfidence, potentially impeding the effective use of financial support in entrepreneurial processes (Hayward et al., 2006).
The study’s implications are significant. Both financial and human capital matter for new business formation worldwide. Our findings suggest that FI policies must account for the decreasing effect in response to high levels of human capital. Education strategies should focus on specific types of education, such as entrepreneurship education with financial literacy, rather than traditional academic curriculum, to foster entrepreneurship knowledge, skills and creativity. Likewise, entrepreneurship support schemes should aim to nurture and share appropriate levels of self-efficacy, avoiding excessively high self-efficacy, which is deleterious to the benefits of FI for entrepreneurial activities.
Despite offering novel insights and practical implications, our research has several limitations that warrant further investigation. First, as with most entrepreneurship studies using GEM data, we were constrained to examining interaction effects on entrepreneurial formation rather than entrepreneurial performance and success metrics, such as revenue growth, profitability or business sustainability. This limitation arises from the nature of the data available. Future research should strive to incorporate measures of entrepreneurial performance and success when analyzing the interplay between FI and human capital, as this would yield deeper insights into strategies for promoting entrepreneurship. Second, the literature suggests that the influence of human capital on the effects of FI may vary by entrepreneurship type – such as opportunity versus necessity, or formal versus informal. However, due to data constraints, this study did not explore these distinctions in relation to the moderating role of human capital. Finally, due to the aggregate nature of our data, we could not explore how FI and human capital influence men and women differently. Given the emphasis on gender equality in the United Nations’ Sustainable Development Goal 5, future studies focusing on how expanded FI benefits women could provide critical guidance for more equitable financial support policies.
