Characteristics of entrepreneurs in the Netherlands: an exploratory study using data from the LISS panel.
Alexander Hartveld University of Amsterdam Faculty of Economics and Business
Abstract
Research about entrepreneurs mainly distinguishes between salaried workers and “entrepreneurs”, usually examined through the self-employed. In our study of
entrepreneurs in the Netherlands, we focus on the incorporated business owners and analyse what characteristics set them apart from salaried workers. We test the
characteristics of entrepreneurs, as proclaimed in the classic theories, by associating them with variables from the representative LISS sample of the Netherlands. Our findings show that two of the Big Five personality traits are negatively correlated with the likelihood to become an entrepreneur: Conscientiousness and Emotional Stability. Openness to Experience shows a strongly positive correlation. Furthermore, we conclude that Need for Cognition is a low-quality predictor of entrepreneurship. These results show more evidence for the entrepreneur in Schumpeter (1934)’s role of the creative leader of the firm, than for Say (1836)’s role of the responsible and capable manager of production. Thus, flexibility is shown to be more important for the entrepreneur than dependability. Our results set the stage for further analysis of specific entrepreneurial traits, family background and country differences.
Characteristics of entrepreneurs in the Netherlands: an exploratory study using data from the LISS panel.
The entrepreneur is at the same time one of the most intriguing and one of the most elusive characters. . . in economic analysis. He has long been recognised as the apex of the hierarchy that determines the behaviour of the firm and thereby bears a heavy responsibility for the vitality of the free enterprise society. – William J Baumol (1968, p. 64)
Introduction
In the past years, there has been a growing number of success stories of tech start-ups in the news, in which individuals sell small innovative companies for
exorbitantly high valuations. For instance Whatsapp, a messaging service company with only 55 employees, was acquired by Facebook for $16 billion in February 2014 (Gelles & Goel, 2014). These headlines have raised one figure back to popular attention, the entrepreneur. However successful these recent stories are, they can hardly be understood as representative for entrepreneurship as a whole; the notion of an entrepreneur has existed way before and beyond the current high-tech boom.
Economists since the 18th century have included the entrepreneur in their analysis as a vital component to economic progress and market development. Also the empirical evidence in a meta-review by van Praag and Versloot (2007) suggests that they generate relatively much employment, productivity growth, and create and commercialize
high-quality innovation. The spill-over effects of entrepreneurship are also known to enhance employment growth rates of other companies in the region over the long run. And finally, the potential for disruptive innovation, as described by Schumpeter (1934), makes entrepreneurs a carrier of technological and economic progress, both in developed as in transitional economies (McMillan & Woodruff, 2003). In short, the entrepreneur has a very important yet specific function in the economy.
In the Netherlands, the entrepreneurial framework conditions are well above the average of other innovation-driven economies according to the Global Entrepreneurship Monitor 2012 (van der Zwan, Hessels, Hoogendoorn, & de Vries, 2012). In terms of the quality of start-ups, Dutch business owners are more pessimistic and significantly less innovative compared to other countries. Nevertheless, Dutch economists and politicians openly praise small and medium-sized enterprises (SMEs) as vital to the Dutch economy (Roth, 2011; Wellink, 2002); Understanding the characteristics of their owners may help
policy makers respond to their needs.
In this paper we explore the characteristics and personal traits of entrepreneurs in the Netherlands. We do so by taking into consideration two recent insights. Firstly, a recent paper by Levine and Rubinstein (2013) points out how a substantial body of entrepreneurship literature fails to distinguish between different types of self-employment,
resulting in biased results about earnings and other observable traits. We believe that the same negligence may have also biased results found in personality-based entrepreneurship literature, where many authors do not distinguish between different types of
self-employment (prominent examples are Brandstätter, 2011; Rauch & Frese, 2007; Zhao, Seibert, & Lumpkin, 2010). We focus on the incorporated self-employed in order to rectify this possible bias. Secondly, the emergence of a “new heterodox mainstream” in the form of bounded rationality and behavioural economics has been particularly useful in conducting research about entrepreneurship (Minniti & Lévesque, 2008). We also believe that our understanding of these economic actors will benefit from including a range of non-cognitive indicators into our analysis, most prominently the Big Five personality traits1.
Firstly, knowing which personality types are most likely to become entrepreneurs may help entrepreneurship research to explain economic outcomes on a micro and behavioural level. Theories may be developed and tested according to our findings about the enterprising individual. Secondly, the results may encourage individuals to follow the challenging but possibly lucrative and innovative career path of venture creation. It may humanize the entrepreneurial figure by presenting an entrepreneurial profile and allow entrepreneurship education to address the opportunities and risks that certain personality traits carry. Thirdly, future research may focus on one specific (personality) trait that stands out in this study. For example Openness to Experience may be investigated further with the goal to understand why this trait is so prevalent among entrepreneurs and what its ramifications are.
Because of the lack of a comprehensive theoretical framework, we will use section 2 to understand the entrepreneur through her2 role in the economy. For this, we draw on the most prominent classic theories in the field3. The roles outlined by those theories suggest certain characteristics of entrepreneurs. In section 3, we then turn towards the LISS dataset from the Netherlands and select the variables that measure these
characteristics best. Our results will be presented and discussed in section 4, followed by a conclusion.
We must make explicit that this paper does not look at entrepreneurial success but rather focuses on the individual that runs an incorporated business, irrespective of its success. The reader should note that the attributes that increase the probability to become an entrepreneur do not necessarily increase the probability of success. Over-confidence, for instance, might be associated with a higher probability of both
1The Big Five personality traits are five broad dimensions captured in a questionnaire that are used to describe human personality. For an overview of the individual traits, please refer to Table 2.
2Although primarily men engage in entrepreneurship, this paper will make equal use of both the female and male pronouns.
3Much of modern entrepreneurship research considers the classic literature relevant in our time, according to van Praag (1999).
engagement in entrepreneurship and business failure.
We will add to the existing literature in three ways. Firstly, to our knowledge we are the first to analyse the characteristics of entrepreneurs as depicted in the classic theories using data from the Netherlands. For this reason, the study will be exploratory and aim to come up with a comprehensive model from the dataset. Secondly, we will include the Need for Cognition measurement into our analysis, an uncommon measure in entrepreneurship research. If the results are conclusive, it can lead to the introduction of a new measure to the toolkit of entrepreneurship research. Finally, this exploratory study contributes to the understanding of the Dutch entrepreneurs and may set the stage for follow-up research or a country profile.
Theoretical Framework
Given the prominent position entrepreneurship takes in modern economic practice, the theoretical framework in this field is remarkably short of a cohesive, explanatory, predictive, or normative theory. Entrepreneurship research still leaves many fundamental questions unanswered and faces the same challenges to theory development as twenty years ago (Amit, Glosten, & Muller, 1993). Firstly, entrepreneurship research is characterized by the interdisciplinary nature of the aspects under investigation.
Notwithstanding the fragmentary demarcation of the field, recent advancements in the field have emerged in a wide array of domains, to cite a few: choice models between alternative employment forms (Lazear, 2004; Levine & Rubinstein, 2013; Parker, 2009; Wagner, 2006), the role of non-pecuniary benefits (Hamilton, 2000; Hurst & Pugsley, 2011), network externalities (see for a critical review Hoang & Antoncic, 2003; Minniti, 2005), personal characteristics (for example Hamilton, Papageorge, & Pande, 2014; Hartog, van Praag, & Van Der Sluis, 2010; Levine & Rubinstein, 2013) and the introduction of new analysis methods (Short, Ketchen, Combs, & Ireland, 2010).
Nonetheless, progression towards a comprehensive theory is not noticeable yet, leaving it up to the researcher to construct a theoretical framework for her own analysis.
The second challenge is that those variables proclaimed to be the causing
differences between e.g. entrepreneurs and salaried workers do not always bear up under scrutiny, as fairly noted in reviews of entrepreneurship literature (for example Brockhaus, 1982; Gartner, 1985). The heterogeneity within the population of entrepreneurs and businesses leaders makes the quest for a homogenous entrepreneurial profile difficult (Gartner, 1985). The definition and demarcation of entrepreneurs will therefore receive
extensive consideration in this study.
Thirdly, some ex ante aspects that might explain entrepreneurial outcomes may be unobservable to outsiders, or similarly the entrepreneur may wrongly identify ex post outcomes as a causal factor (Amit et al., 1993). For our quantitative analysis this means that the traits we find in entrepreneurs are only estimated correlations, which may
actually result from an unobserved third variable that affects both the analysed trait and the likelihood to become an entrepreneur.
Finally, the prevalence of the neo-classical paradigm with its notions of perfect information and rational choice has left no room for an active entrepreneur (Baumol, 1968; Busenitz et al., 2003; van Praag, 1999). Despite the loosening of some assumptions due to critique4, the vast majority of research still considers the entrepreneur to be perfectly aware of his entrepreneurial capacities and expected returns to entrepreneurship. In light of the widely acknowledged influence of uncertainty and luck in the success of entrepreneurship, this seems to be counter-intuitive. The currently perceivable
progression towards a heterodox mainstream in the form of bounded rationality and behavioural economics therefore facilitates entrepreneurship research.
Many questions are still open in entrepreneurship research, and still today attempts to answer have by far not been exhaustive or consolidating towards a theory. As Amit et al. (1993, p. 831) noted, it may “be somewhat too ambitious to expect a complete theory of entrepreneurship.” The modern notions of bounded rationality and cognition are shifting the paradigm away from the neoclassical orthodoxy, benefiting our type of research.
The open questions and disagreement about the right theoretical framework make the classic theories a popular point of departure for researchers in this field. This is observable as in that most current contributions to the discipline still refer to the
theoretical fathers like Schumpeter, Knight and Leibenstein. Backed by current empirical results, van Praag (1999, p. 334) suggests that the components of these theories still hold in our time and are considered “extremely valuable and applicable”. The role of the entrepreneurs to the economy has not changed much; one can even argue that their relevance has increased5. It is for this reason that we will proceed with a brief review of the main entrepreneurship theories to obtain testable characteristics, as summarized in Table 16. In the section on research design we will link these characteristics to the variables provided in our dataset, creating a comprehensive and testable model.
The role of the entrepreneur in the economy
Bearing of Risk. Following F. H. Knight (1921), the entrepreneur is the
uncertainty-bearing agent in a market characterized by limited information and fluctuating demand, prices, and thus profits. He guarantees the other factors of
4Kihlstrom and Laffont (1979) for example, use differences in risk aversion within the neo-classical school in order to explain the existence of the entrepreneur within the perfect rationality and perfect information assumptions.
5One can observe how for example Schumpeter’s notion of disruptive technology (1934) has become more popular in our time of rapid technological advancements.
6van Praag (1999) gives a more extensive overview of classical theories while comparing them to empirical results, and Long (1983) investigates the definition of entrepreneurship according the classic literature.
production a fixed compensation, thereby assuming the uninsurable business hazard. The entrepreneur uses his judgment to identify profitable opportunities and to properly estimate the required input and output prices, thereby gaining an according profit or loss. The personal characteristics that this role implies are comfort with uncertainty, a high degree of self-confidence, a venturesome nature, and judgement coming from intelligence and experience (F. H. Knight, 1921).
Capable coordination of production and carrying responsibility. Almost a century
before Knight, Jean-Baptiste Say (1836) attributed the task of combining and coordinating the factors of production to the entrepreneur, allowing her to take the residual as profit. She is the modern leader and manager within her firm. In the 20th century, Leibenstein (1968) comes up with a similar role, adding to it the carrying of the ultimate responsibility for the organization and its external obligations. Say (1836, pp.330-331) describes the unique characteristics as “a combination of moral qualities that are not often found together. Judgment, perseverance, and knowledge of the world as well as of business . . . the art of superintendence and administration”. Furthermore, this role requires the entrepreneur to be communicative, responsible, and willing to take control, while acting with perseverance, confidence, and reliance on experience.
Gap and Input Filling. The second major element of the contribution to the topic
by Leibenstein (1968) is the notion of gap-filling as an entrepreneur’s key function. This means firstly that in case problems arise within the firm, it is the role of the entrepreneur to solve them. Second, the entrepreneur also has to provide certain inputs that are by nature vague and for which there is no clear-cut output, for example motivation and leadership. Thirdly, in case six inputs are needed to create a marketable product, marshalling five is not enough and the entrepreneur will have to fill in with her own ability. These points require the following characteristics of an entrepreneur: flexibility, comfort with unfamiliar situations, leadership, emotional stability, and being a
jack-of-all-trades as opposed to a specialist7.
Creative Leadership. Arguably, one of the most noteworthy contributors to the
theory of entrepreneurship was Joseph Schumpeter (1883-1950). He turned down the notion of the entrepreneur as manager of the firm and instead considered his role as the leader of the firm with the denotation used in modern business language. According to Schumpeter (1934, 1939) entrepreneurs are the driving force of progress in the economic system, where their innovation disrupts the equilibrium in the economy and brings change and improvements. Accordingly, the capability to move an organization beyond the current equilibrium requires a rare attitude and a particular conduct. This role manifests itself in characteristics such as leadership, self-confidence, the mental freedom and creativity to do something new, and strength and flexibility need to challenge the
7Lazear (2004) built his popular theory upon this, supported by evidence from German and American data (Wagner, 2003).
current equilibrium.
Procuring Financing. Whereas Schumpeter (1939) explicitly excludes the burden of
financial risk and pressure from the business of the entrepreneur, credit constraints are an element contained in most other theories. The exact effect of capital constraints is disputed in the theories, yet a case can be made that entrepreneurs must secure financing for their ventures. Some authors such as F. H. Knight (1921) and Leibenstein (1968) argue that due to uncertainty and asymmetric information, the entrepreneur needs to have access to personal resources; a requirement that is empirically supported (see for example Evans & Jovanovic, 1989; Marlow & Patton, 2005). Banks, venture capitalists, and other capital providers generally require entrepreneur to provide personal funds from savings or other sources such as friends or relatives in order to invest in unproven companies (Parker, 2009, pp. 203-211). Characteristics therefore include the direct or indirect access to capital, through personal savings, family wealth or a wealthy network, and possibly some familiarity with economic and business thinking paired with
self-confidence to convince other investors.
Now that the classic theories have laid out their understanding of the roles of the entrepreneur8, we can proceed with the theory that links these roles to our empirical analysis. We found career choice theory (Holland, 1997; Lent, Brown, & Hackett, 1994) complemented by the person-environment fit theory (Judge & Kristof-Brown, 2004; Kristof-Brown, Zimmerman, & Johnson, 2005) to be the most applicable decision framework. Career choice theory highlights the means by which individuals exercise personal agency in their career development and is based on social cognitive theory. Person-environment fit theory considers the link between the individual’s characteristics and the job-, organization-, supervisor- and group environments. Together, they form a social cognitive theory of career choice, which suggests that individuals choose a work environment that matches their personality, values, needs, and interests. F. H. Knight (1921); Leibenstein (1968); Say (1836); Schumpeter (1939) all argue that the challenging
role of an entrepreneur requires a unique set of characteristics. The person-environment fit theory explains from the perspective of individuals why they might feel attracted to an entrepreneurial career. Consistent with the person-environment fit and career choice theory, we expect that people who perceive a match between their own personality and the roles of an entrepreneur will self-select towards this occupation9. Before we can proceed though, we must operationalize the central figure of our analysis, the entrepreneur.
8Long (1983) argues that modern definitions of entrepreneurship, which exclude any of the following three fundamental dimensions, are basically incomplete: 1) uncertainty and risk, 2) complementary managerial competence, and 3) creative opportunism.
9This approach has proven to be valuable in other empirical literature (Zhao et al., 2010), and we expect it to contribute to our understanding of the entrepreneurial figure in the Netherlands.
Table 1
Roles and Characteristics of the Entrepreneur
Role of the entrepreneur Characteristics
Bearing of risk (F. H. Knight, 1921)
Comfort with Uncertainty, Self-Confidence,
Venturesome Nature,
Judgment, Intelligence, Foresight.
Capable coordinator of production and carrying responsibility (Leibenstein, 1968; Say, 1836)
Business Administration Expertise, Responsible,
Good Communication Skills, Judgment, Intelligence, Foresight, Experience (knowledge of the world), Perseverance,
Self-Confidence.
Gap and Input Filling Capacity (Lazear, 2004; Leibenstein, 1968)
Flexibility,
Broadly skilled and experienced, Leadership,
Emotional Stability,
Motivation to solve problems.
Creative Leadership (Schumpeter, 1939)
Leadership, Self-Confidence,
Mental Freedom to try something new, Creativity,
Flexibility and Adaptability.
Procuring Financing (F. H. Knight, 1921;
Leibenstein, 1968)
Familiarity with Business and Finance, High Personal Savings,
Wealthy family or network, Self-Confidence.
Demarcating the entrepreneur
What should be a substantial consideration in entrepreneurship research is astonishingly lightly taken in current research: the demarcation of the entrepreneur. Even principal contributors to the domain do away with hasty definitions of “the
self-employed” or “those that own a business” while the ambiguity associated with these terms is substantial. Gartner (1985) accurately recognizes the heterogeneity within the entrepreneur population as a validity problem for drawing conclusions about an assumed homogenous entrepreneur. We argue in line with Levine and Rubinstein (2013) that the self-employed are not a good proxy for growth-creating innovation and follow their precedent by studying incorporated businesses.
Incorporation has been created to facilitate entrepreneurship through two defining characteristics: limited liability and a separate legal identity. Limited Liability protects
its owners from huge downside losses, increasing the appeal of engaging in innovative but risky projects. The separate legal identity in turn protects the firm from owner-specific shocks and increases the appeal of investing in long-term development projects. At the same time, the larger direct and indirect costs associated with the incorporated form furthermore make it unattractive for those without relatively high stakes or expectations (Fama & Jensen, 1985; Horvath & Woywode, 2005). We can therefore reasonably assume
that the entrepreneur from our classical theorists would nowadays have founded an incorporated business.
Research Design
We have set up the appropriate theoretical framework about the entrepreneur, by describing her role in the economy, her assumed person-environment fit decision model, and in what business forms we should look for her in empirical research. In order to proceed with our analysis of the Netherlands, we must operationalize the theory
according to the variables in the LISS dataset. In this part, we describe the design of our analysis and what variables precisely we will use as a proxy for each characteristic of an entrepreneur. We will select a range of non-cognitive, demographic, and family
background variables for this purpose.
Methodology
In order to investigate the effect of non-cognitive traits, and (family) background in the Netherlands on the likelihood of being an entrepreneur, we estimate by Maximum Likelihood Estimation (MLE) the following probability model:
P r (Yi = yi | Xi) = 1 1+e−F) yi 1 − 1+e1−F) 1−yi Where F = β · Xi
The role played by the explanatory variables in determining entrepreneurship is investigated in this logistic regression model with the dependent variable Y taking the value 1 if the person was an entrepreneur at some point during the survey, and 0
otherwise. Furthermore, β represents the vector of unknown parameters to be estimated, and Xi represents the vector of the explanatory variables, which are categorical, binary,
and continuous.
For our purposes we will test four models with a modified set of variables each as a means of controlling for conditional dependency of the variables. The base model only includes the demographic and the control variables gender, age, education and whether the person has a partner. The second model includes the personality measures as described below, whereas the third model includes the family background. The fourth model is a comprehensive model including both psychological measures as well as
information about the family background. The included variables are described in the following subsections and can be split into three categories: demographics, non-cognitive traits, and family background.
Sample
The data used in this study comes from the LISS panel (Longitudinal Internet Studies for the Social Sciences), administered by CentERdata (Tilburg University, The Netherlands). It comprises a representative sample of approximately 5000 Dutch
households or 8000 individuals who participate in monthly internet surveys and receive a compensation for their participation (Scherpenzeel & Das, 2010). Starting in 2007, the panel fields two types of studies: the LISS Core Studies, which are yearly longitudinal studies, and several single-wave studies on fields that are of interest to selected social scientists or political institutions. Its variables cover a large variety of domains including work, education, income, housing, time use, political views, values and personality.
The broad range of personality indicators and the differentiation between
alternative forms of self-employment makes it the most suitable secondary data available for the Netherlands for our type of research. In our study we assemble responses from the following questionnaires:
• Background variables LISS Panel (LISS Core Study waves 2007-2013) • Economic Situation: Income (LISS Core Study, waves 2007-2013) • Personality Core Study (LISS Core Study, waves 2007-2013) • Life History Questionnaire (LISS Assembled Study, wave 2012)
When comparing entrepreneurs to salaried workers, we need to perform a few corrections to our sample. Firstly, to select persons with a realistic perspective of starting a business, we excluded those that are younger than 18 years and those that are older than 65 years from our analysis. Secondly, our sample of salaried workers excludes those that follow a full-time study, take care of housekeeping, are pensioners, have a work disability, work as volunteers, or who indicated to “do something else”. We do keep students in our sample of company owners to allow for student entrepreneurs10. And thirdly, we excluded those that had not or only partially completed the personality tests. One outlier who had indicated a monthly income ofe 233.300 was also removed from the dataset.
In order to qualify for our definition of an entrepreneur, our subjects must have indicated in at least one of the six questionnaire waves to be the owner of a limited
10Our sample of 214 entrepreneurs contains 12 students that report to own an incorporated business next to their studies.
liability company (in Dutch: Commanditaire Vennootschap or Besloten Vennootschap) or have indicated to be the owner of a company with people employed. Because of the short time span in which data was collected (the LISS panel has information for six years only), we employ the data in a cross-sectional analysis instead of in a time-series analysis.
We have listed the summary statistics of incorporated business owners, owners of a one-person business and partners of a commercial partnership separately in Table A2, giving us evidence that it was justified to focus on the incorporated business owners for our understanding of the entrepreneur. The control group only contains salaried workers.
Every month, participants of the LISS panel check for changes in their background data, resulting in a large number of waves per person with information on income, education, age, gender, civil status, etc. In a cross-sectional analysis we only can use data from one wave, so for the self-employment group, we used the most recent year in which they owned a business. For the salaried group, we chose the year in which the median value of the Need for Cognition test was obtained. If the median was rounded and cannot be allocated directly, we use the last year the participant provided a personality test.
For the range of included psychology measures, we calculated the test score and standardized it to be situated between zero and one. We did not use a dummy variable for their measurement because of the loss in variance this would result in and the arbitrarily chosen cut-off point. However, the interpretation of the corresponding estimators is troublesome due to the conditionality of the results, and we can use their value only as an indication about the strength of the effect. Furthermore, the panel repeats the same personality tests every year and the median scores of each personality indicator were included into the model.
Explanatory Variables
The explanatory variables that we include can be divided into three categories: Background Variables (Demographics), Non-Cognitive Traits, and Family Variables. In this section, each variable is described and its inclusion is motivated through one or more of the entrepreneurial roles from the classic theory. We give an overview of the variables included and how they relate to the characteristics from the five roles in Table 3, with an indicator of the evidence we found.
Background Variables. The personal background variables we include are gender, a quadratic term of age, highest education that was achieved with a diploma and whether this education was in the field of economics or business. Since the LISS panel unfortunately includes no measure of working history, the closest indicator for experience that we can include is age. The expected positive effect of experience could be both reduced or reinforced by the effect age has, which can arguably affect entrepreneurial intentions in both directions (see Parker, 2009, p. 113 for an elaboration on the effect of age, with references to empirical evidence for both direction.). Experience is most
important to the role of the coordinator of production, as described by Leibenstein (1968) and Say (1836). Furthermore, higher personal savings usually come with a higher age, which is important for the role of the entrepreneur as the person with access to financing.
Hypothesis 1a: According to the roles of coordinating production and procuring financing, age has a positive effect on the likelihood of someone being an entrepreneur.
Formal education has been included in the form of dummy variables for six
education categories. As with age, one can argue for both directions of the effect of formal education; on one hand, education might improve entrepreneurial judgment, foresight, and imagination, while on the other hand it increases e.g. the value of the outside option of paid employment. Additionally, skills that make entrepreneurs successful are usually not found inside the curriculum. Therefore, we cannot expect a priori that either general or specialized education should be related to entrepreneurship11. However, both theory and empirical findings suggest that an education in the field of business administration or economics will have a positive effect. Especially to the role of the entrepreneur as a coordinator of production this is important; For example accounting and market research skills will help the entrepreneur lead the production and recognize opportunities.
Secondly, for the role of procuring financing, we also argued that an understanding in this field might make it easier to convince financiers of the enterprise’s success.
Hypothesis 1b: According to the roles of coordinating production and procuring financing, education in a economics or business related has a positive effect on entrepreneurship.
Non-Cognitive Traits. Personality theory regained territory in
entrepreneurship research, after being disregarded for almost 20 years. By the late 1980s namely, narrative reviews of literature on personality had concluded that the relationship of personality to entrepreneurship outcomes lacked consistency (see Baumol, 1968; Brockhaus, 1982; Gartner, 1988). Only now in more recent years, multiple meta-analyses demonstrated that the empirical findings are far more consistent than suggested by the narrative reviews (Rauch & Frese, 2007; Zhao et al., 2010)12. As we include the
personality traits into the model, it is important to reinforce our proposition that individuals are attracted to roles that match their personality and interests. This proposition follows the person-environment fit theory as proposed in section 2.
The pronounced return of personality theory to entrepreneurship theory came around the same time as the consensus that the basic structure of human personality can be captured by five broad personality dimensions, namely Emotional Stability,
Extraversion, Openness to Experience, Agreeableness, and Conscientiousness; also known as the Big Five and summarized in Table 2. Most important for their use in economic
11For a more detailed analysis with empirical results, see papers referred to in Parker (2009, p. 117) 12The meta-analyses by Brandstätter (2011); Rauch and Frese (2007); Zhao et al. (2010), however, do not adequately make distinctions between types of the self-employed used in their sources. This is a possible source of confusion and bias, as noted earlier.
research is that they appear to be stable over the lifecycle (Cobb-Clark & Schurer, 2012), reducing concerns about simultaneous causality.
The first trait, Extraversion, characterizes people as outgoing, warm, and friendly. They are energetic, assertive, and dominant in social interactions, and experience more positive emotions and optimism. Extraversion is not necessarily a prerequisite for leading a business, however its attributes are generally associated with people’s implicit
perception of the leadership role (Lord, De Vader, & Alliger, 1986). Therefore it will likely help coordinating the work of others, as in the entrepreneurship model by Leibenstein (1968). Also in the model where the entrepreneur procures access to financing,
extraversion may be beneficial. One must be able to approach possible investors within the direct environment, use (in)-formal networks, or convince bankers to lend money.
Hypothesis 2a: According to the roles of coordinating of production, gap- and input filling, and procuring financing, Extraversion has a positive effect on entrepreneurship.
Conscientiousness describes an individual’s level of dependability; that is being
responsible, careful, thorough, organized, and planful. It is also linked to a need for achievement, work motivation, acceptance of traditional norms and virtues, and
responsibility towards others (Barrick & Mount, 1991; Costa Jr & McCrae, 1992). The predicted effect differs between the classic theories of the entrepreneurial role as a coordinator of production (as described in Say (1836) and Leibenstein (1968)), and the role of leading the firm towards innovation (as defined by Schumpeter, 1939). Say’s persistence, responsibility and hard work seem to stand opposite in this trait towards the mental freedom and creativity that Schumpeter praises in the entrepreneur. We will thus formulate two conflicting hypotheses.
Hypothesis 2b: According to the role of the entrepreneur as a capable and responsible coordinator of production, Conscientiousness has a positive effect on entrepreneurship.
Hypothesis 2c: According to the role of creative leadership, Conscientiousness has a negative effect on entrepreneurship.
Openness to Experience is a personality trait assigned to those who are
intellectually curious, imaginative, and creative; someone who seeks out new ideas and alternative values. All theories concur in judgment about this trait; it can be related to comfort with uncertainty, flexibility, intelligence, mental freedom, opportunity recognition and the characteristics of creative leadership. Ciavarella, Buchholtz, Riordan, Gatewood, and Stokes (2004) and Morrison (1997)) link it to opportunity recognition, and the meta-analysis by (Zhao et al., 2010) supports the notion with empirical evidence. We can see the strongest ties to the role of the creative leader (the role proposed by Schumpeter, 1939), the bearing of risk (F. H. Knight, 1921), and the gap- and input filling role, where flexibility and openness to different tasks and problems are important.
Hypothesis 2d: According to the roles of bearing risk, gap- and input filling, and creative leadership, Openness to Experience has a positive effect on entrepreneurship.
People that score high on Emotional Stability are characterised as well-balanced, stable, even tempered. On the other hand, those that are low on emotional stability (also referred to as Neuroticism) are more vulnerable to stress, sensitive to negative feedback, and experience negative emotions more intensely; foremost anxiety and worry. In the roles the classic theory describes, entrepreneurs carry the responsibility (Say, 1836) and uncertainty (F. H. Knight, 1921) and concern themselves with solving problems and making tough decisions as they arise (Leibenstein, 1968).
Hypothesis 2e: According to the roles of bearing risk, coordination of production and input- and gap filling, Emotional Stability has a positive effect on entrepreneurship.
The fifth and last personality dimension is Agreeableness, which indicates the stance on personal contact with others. Those scoring high on the agreeableness scale are considered trusting, altruistic, cooperative and prefer consensus, while those on the other side of the spectrum are characterized as manipulative, self-centred, suspicious, and ruthless. Depending on the tasks and role of an entrepreneur, both sides of the spectrum can be of relevance to the responsibilities of the entrepreneur. More in line with
Schumpeter’s entrepreneurial figure we can argue that for creativity to thrive, one needs to be trusting and cooperative and ideas should be shared (Martins & Terblanche, 2003). The other side of the classic theory sees the entrepreneur more as the manager of
production and bearer of uncertainty. In this role, the business owner might have to deal with strong competition, stand firm against suppliers, and make others work for their personal needs and interests (as described by Singh & DeNoble, 2003). The empirical evidence from Zhao et al. (2010) also indicates a negative relationship.
Hypothesis 2f: According to the role of creative leadership, agreeableness has a positive effect on entrepreneurship.
Hypothesis 2g: According to the roles of bearing risk and the coordination of production, Agreeableness has a negative effect on entrepreneurship.
Table (2) The Big Five Personality Traits
Personality Trait: Associated With:
Extraversion Persons that are outgoing, warm, friendly, ener-getic, assertive, and dominant in social interac-tions.
Conscientiousness Dependability, good impulse control, goal- directed behaviors, good organization and mindfulness of details.
Openness to Experience Imagination, intellectually curious, creative, seek-ing out new ideas and alternative values.
Emotional Stability Well-balanced, emotionally stable, even tempered, calm, and free from persistent negative feelings Agreeableness Trust, altruism, kindness, affection, and other
pro-social behaviors
Notes: Description of each personality trait, adapted from (Hamilton et al., 2014)
We further use the Rosenberg Self-Esteem score, which measures the degree of approval of one’s self. It is a ten-part questionnaire included in the personality wave of our dataset and has been widely used in psychology and economics (for example in Bowles, Gintis, & Osborne, 2001; Heckman, Stixrud, & Urzua, 2006). It measures how much the entrepreneur believes in his own ability and therefore may be more inclined to leave their current occupation to start up their own business. Also, supported by
empirical findings (Levine & Rubinstein, 2013), someone with a high self-esteem will be more likely to believe in their own ability to handle risky situations, and their superior knowledge on moving the firm beyond the current equilibrium (as deemed important by Schumpeter, 1939).
Hypothesis 2h: According to the roles of bearing risk, gap- and input filling, creative leadership, and procuring financing, the Rosenberg Self-Esteem score has a positive effect on entrepreneurship
We also decided to include Need for Cognition, so far an uncommon measure in entrepreneurship research. It captures the “tendency for an individual to engage in and enjoy thinking” and was developed by Cacioppo, Petty, and Feng Kao (1984). Individuals rate to which extent they agree with each of 18 statements such as “I find satisfaction in deliberating hard and for long hours,” “The notion of thinking abstractly is appealing to me,” and “Thinking is not my idea of fun” (Cacioppo et al., 1984). The reasoning to include it13 is that in the role of a bearer of risk or as a capable coordinator of inputs, the
entrepreneur will be faced with difficult decisions. To handle this sensibly, she will need to rely on her judgment and foresight resulting from deliberate reflection. We argue that self-selection into entrepreneurship is more likely for those that enjoy reflecting about problems, something that may also be in line with Schumpeter (1939)’s notion of creative leadership. Because the reliability and values of this measure are still unknown, Table A4 also presents the regression results without the Need for Cognition measure. Depending on the results we might be able to include this measure into the toolkit of an
entrepreneur.
Hypothesis 2i: Need for cognition is a reliable and valuable tool that can be added to the toolkit of an entrepreneur.
Hypothesis 2j: According to the roles of bearing risk, and creative leadership, Need for Cognition has a positive effect on entrepreneurship.
Family Background
The influence of parental education on the entrepreneurship choice is a prevalent factor to control for in entrepreneurship research. Although it is not directly related to any of the entrepreneurship theories we employ, parental education has been
demonstrated to have a strong influence on the likeliness to become an entrepreneur (see Parker, 2009, pp. 134-138). For this purpose and consistent with the approach by Levine and Rubinstein (2013), we calculate the years of education using averages from the CBS Statistics Netherlands data for different types of education(Elbers, 2009).
Hypothesis 3a: In line with empirical findings, we expect parental education to have a positive effect on entrepreneurial intention.
In order to include the access to finance into our analysis, we have only access to a question about Family Wealth: “How would you describe the material situation of the family in which you lived at the time?”. We use a dummy variable if the answer is “very prosperous” or “fairly prosperous”, the two highest of five options. Although self-reported personal assets are also available in the dataset, it was not included to prevent inverse causality associated with personal wealth and the success of entrepreneurship. Family Wealth is relatively independent of personal success while also making access to financing more likely; possibly via access to wealth in the family or via the resulting network14. All classic theorists, with the exception of Schumpeter, agree about the importance of procuring financing. And indeed, also the empirical literature has documented this relationship between initial wealth and subsequent business entry to be positive (see e.g.
unfortunately missing in the LISS data (Cacioppo et al., 1984).
14However, only 6-8 per cent of business owners borrow capital from family members according to Aldrich, Renzulli, and Langton (1998); Fairlie and Robb (2007). R. M. Knight (1989) reported that the following sources of funds were used by high-tech Canadian forms at the pre start-up stage: personal savings: 60 per cent; family and friends: 13 per cent; bank loans: 12 per cent; and trade credit: 6 per cent. The private information family members might have about borrowers may explain their motivation as a source of financing.
Evans & Leighton, 1989). Furthermore, higher family wealth is shown to also increase financial optimism, which in turn is also important for the willingness to bear risk (Dawson, De Meza, Henley, & Arabsheibani, 2012). We see that (family) wealth variable
is most important for the role of procuring financing for business ventures.
Hypothesis 3b: Family wealth when growing up will have a positive effect on entrepreneurship, considering the importance access to resources for entrepreneurs.
Table 3 gives an overview of all hypotheses and how they relate to the
characteristics proposed by the classic theories, including the evidence that we will discuss in the next section on results.
Table (3) Roles of the Entrepreneur and Included Variables
Characteristics Variables in the model
Bearing of Risk (F. H. Knight, 1921)
Comfort with Uncertainty, Self-Confidence,
Venturesome Nature,
Judgment, Intelligence, Foresight.
Emotional Stabilityrejected,
Openness to Experience***, Agreeableness (neg.),
Self-Esteem (Rosenberg)*, Need for Cognition.
Capable coordinator of production and carrying responsibility (Leibenstein, 1968; Say, 1836)
Business Administration Expertise, Responsible,
Good Communication Skills, Judgment, Intelligence, Foresight, Experience (knowledge of the world), Perseverance,
Self-Confidence.
Age**,
Education in Economics or Business ***, Extraversion,
Conscientiousness (pos.)rejected, Emotional Stabilityrejected, Agreeableness (neg.), Self-Esteem (Rosenberg)*.
Gap and Input Filling Capacity (Lazear, 2004; Leibenstein, 1968)
Flexibility,
Broadly skilled and experienced, Leadership,
Emotional Stability,
Motivation to solve problems.
Age**, Extraversion,
Openness to Experience***, Conscientiousness (neg.)**, Emotional Stabilityrejected,
Self-Esteem (Rosenberg)*.
Creative Leadership (Schumpeter, 1939)
Leadership, Self-Confidence,
Mental Freedom to try something new, Creativity,
Flexibility and Adaptability.
Openness to Experience***, Self-Esteem (Rosenberg)*, Conscientiousness (neg.)**, Agreeableness (pos.), Need for Cognition.
Procuring Financing (F. H. Knight, 1921; Leibenstein, 1968)
Familiarity with Business and Finance, High Personal Savings,
Wealthy family or network, Self-Confidence
Education in Economics or Business***, Age**,
Family Wealth,
Self-Esteem (Rosenberg)*.
Notes: Significance Levels:* p<0.1, ** p<0.05, *** p<0.01. Rejected means that the hypothesis was rejected due
Results and Discussion
The first part of this section presents the summary measures of our sample, showing differences between the different categories of self-employed. Before proceeding to a presentation of results and their respective discussion, we evaluate the correlations and validity of the psychological measures. We then discuss the evidence per hypothesis as summarized in Table 5. Next, we compare the results in the light of the classic theories we drew from, with an overview given in Table 3. Finally, a discussion about the validity and limitations of our results will aid us in drawing the appropriate conclusions about the results.
Table A2 shows a number of summary measures about the constituents of our sample: Salaried workers, incorporated business owners, partners in a commercial partnership, and single-person business owners. We firstly notice that among the incorporated self-employed, the percentage of males, university graduates, and people with an education in economics or business is significantly higher than for the other classes (18.70 % compared to 10,68%). Secondly, the amount of working hours per week by the incorporated self-employed is substantially higher than in the other groups. Thirdly, also earnings appear to be significantly higher for the incorporated self-employed, yet this figures should be evaluated with caution. Self-reported net income is prone to underreporting or misinterpreting of questions (Knoef & de Vos, 2009), e.g. around 10% of incorporated entrepreneurs indicated not to know their income. As Evans and Leighton (1989) note, self-employed might report a fixed salary, but can also include or exclude
entrepreneurship profits and returns, depending on the interpretation of the question and the features of the business. Fast growing businesses, for example, might reinvest their earnings, while a business liquidation might lead to large cash flows. Notwithstanding this limitation, we do have an indication that income is significantly higher for incorporated self-employed, something that is in line with empirical findings by Levine and Rubinstein (2013). In conclusion, these vast differences in summary statistics strongly confirm the
importance of distinguishing between the various types of self-employed businesses, illustrating the gravity of the omission by much of the empirical literature.
When we zoom in on the incorporated self-employed, we see in Table 4 that in total 58 of 221 business owner (26,24%) have abandoned their incorporated self-employment during the period of study. Almost half of them did so during the year 2008, likely as a result of the financial distress from that period. For our research topic, the characteristics of entrepreneurs in the Netherlands, this exogenous shock is less of an issue than were we to study the success factors of entrepreneurs15. Furthermore it should be noted that 71 of 15We tested the effect of individual waves by the inclusion of a dummy variable for each year the background data was taken from. The results showed no noteworthy effect and therefore these measures were excluded.
221 business owners dropped out of the panel due to attrition16. Table 4
Attrition of Incorporated Self-Employed
¸Wave ¸Cumulative N of Owners
¸Observations in wave
¸Attrition ¸Quit their Busi-ness 2008 98 106 14 6 2009 120 86 15 28 2010 158 81 13 12 2011 169 68 9 4 2012 198 86 20 8 2013 214 71 Total 214 71 58
Upon inspection of the correlation table of the personality tests(presented in Table A1, two values stand out. Firstly, the Rosenberg Self-Esteem coefficient and the coefficient for Emotional Stability correlate to a high extent (r=.615). This relationship has repeatedly been observed in other studies and Quilty, Oakman, and Risko (2006) attribute its existence to an increased likelihood that those high on the Neuroticism scale will endorse negatively worded statements. Although influenced by it, Self-Esteem still measures a characteristic distinct from the Big Five and both variables show significant results. The second value that catches the eye is the correlation between Openness for Experience and the Need for Cognition. Sadowski and Cogburn (1997) explain that this relationship seems to reflect the joint measurement of enjoyment to engage in cognitive activity. A separate regression was run without the Need for Cognition measure in Table A4 and, apart from a lower (but equally significant) Openness to Experience coefficient, we observe little impact on the results. In summary, the Rosenberg’s
Self-Esteem score is said to keep its construct validity under correlation with Emotional Stability, and Openness to Experience shows similar statistical evidence and a lower coefficient a Need for Cognition parameter in the model. It is therefore safe to assume that the before mentioned correlation will not pose a large threat to the validity and at most lead to inflated standard error measurements.
The results from the regression, presented in Table A3 (including Need for
Cognition and Table A4 (excluding Need for Cognition), allow us a broad insight into the workings of entrepreneurs in the Netherlands. When comparing the models, we see that both family background and psychological factors substantially add to the explanatory power of the model as measured by the pseudo R2 measure. When all factors are
16One might reason that attrition is higher for entrepreneurs as they less of an incentive to fill in the questionnaires due to longer working hours and higher opportunity costs. And indeed, when comparing attrition rates between self-employed and salaried workers we note that attrition of the self-employed is four percentage points higher; a negligible difference (De Vos, 2009).
included, the fit is the highest and also the coefficients change; controlling for family background increases the effect size of multiple personality measures and makes Emotional Stability and Conscientiousness significant while achieving the opposite for Extraversion. At the same time, the explanatory power of age (or perhaps experience) decreases due to the included covariates. For the interpretation of the results, it should be considered that the coefficients in the model conditional and the reported values assume that all other variables are being kept constant at the mean.
The conditional effect sizes of variables in the
com-prehensive model (Model 4) should be considered when interpreting the following variables.
Table (5) Hypotheses
Relationship to the likelihood of someone being an entrepreneur:
Result:
1. Demographics
1a Age has a positive effect Significant Evidence** 1b Economics or business as the field of
education have a positive effect
Significant Evidence***
2. Non-Cognitive Traits
2a Extraversion has a positive effect No Significant Evidence 2b Conscientiousness has a positive effect Rejected
2c Conscientiousness has a negative effect Significant Evidence** 2d Openness to Experience has a positive effect Significant Evidence***
2e Emotional Stability has a positive effect Rejected 2f Agreeableness has a positive effect No Evidence
2g Agreeableness has a negative effect No Significant Evidence 2h Self-Esteem has a positive effect Significant Evidence*
2i Need for Cognition is a reliable and valuable tool in the analysis of entrepreneurship
Rejected 2j Need for Cognition has a positive effect No Evidence
3. Family Background
3a Parental Education has a positive effect Significant Evidence for father, no evidence for mother
3b Family Wealth has a positive effect No Significant Evidence Notes: Significance Levels:* p<0.1, ** p<0.05, *** p<0.01. Hypothesis rejected means that there is
evidence for an effect in the opposite direction. No evidence means that an effect in the opposite direction is observed, although not significant.
Background Variables
Entrepreneurship is still an activity primarily for male actors, as the 75% rate of males suggests. This is by far the highest rate compared to the other forms of
self-employment, being 54% for partners and 48% for the single-person businesses. We should bear in mind here that despite the low percentage of women, this observation gives no indication about the success of female entrepreneurs, as this is not measured in our sample.
Significant evidence for Hypothesis 1a. The likelihood of entrepreneurship increases
exponentially with age, showing significant coefficients depending on the model. The age squared is significant at the 10% level in our comprehensive model and significant at the 1% level in the base model and the model including personality measures only. Figure 1 illustrates the age distribution of entrepreneurs in our sample. The empirical studies summarized by Levesque and Minniti (2006) show a similar distribution with
entrepreneurship being concentrated around individuals in mid-career, meaning between 35 and 45 years of age. In our sample, we observe a large jump in the number of
entrepreneurs around ten years earlier, followed by stabilization. The distribution likely flattens before falling at age 60 as those that were successful their business will likely stay with their business17. Future research could be directed to the reasons for this earlier age of starting a business in the Netherlands.
Figure 1 . Age Distribution of Incorporated Business Owners
0 .01 .02 .03 .04 Density 20 30 40 50 60 Age
Significant evidence for Hypothesis 1b. Our results show that the field of education
matters and not the type of the diploma earned. While the type of diploma has no observable effect, a diploma in the field of economics or business significantly increases
the chance of becoming an entrepreneur. Table A4 shows that when controlling for psychological factors, the effect becomes even larger. This can be interpreted as that for individuals with the correct personality profile and family background, the marginal effect of education in this field is substantial. Also, this finding is in line with the statement by Say (1836) that an entrepreneur requires knowledge “of business . . . the art of
superintendence and administration”. Finally, this finding can point to the success of business education in the Netherlands, in accordance with the appraisal by the Global Entrepreneurship Monitor (van der Zwan et al., 2012).
Significant evidence for Hypothesis 2d. From the Big Five Personality traits, the
most significant characteristic (p=0.002) of an entrepreneur is the Openness to
Experience; both the significance level and the relative size of this effect stand out in our results. We argued that this could be due to an increased willingness to engage in the uncertainty of starting a venture. The second explanation is derived from higher imagination and creativity that is associated with this trait (Brandstätter, 2011),
supporting the role from Schumpeter (1934, 1939). In their empirical analysis of the Big Five personality traits and their effect of job performance, Barrick and Mount (1991) found no conventional occupational group that scored particularly high on the Openness to Experience scale. We now have evidence that the entrepreneur might have been missing from their analysis. This trait can also give a causal explanation why illicit behaviour was such a significant predictor in entrepreneurial intentions in the paper by Levine and Rubinstein (2013). Those that are more open to experience might also be more likely get utility from activities as speeding, experimenting with drugs, and the thrill of other illicit behaviour.
Hypothesis 2e is rejected. Contrary to all predictions and most empirical findings (Zhao et al., 2010), our coefficient for Emotional Stability is significantly negative
(p=0.033). This provides evidence that those scoring low on Emotional Stability scale are more likely to become incorporated business owners in the Netherlands. The
characteristics of those scoring high on Neuroticism (the inverse of Emotional Stability) are that they experience anger, anxiety, depression and vulnerability stronger. This result requires further investigation, possibly in consideration of the Person-Environment Fit theory. One can argue that those indicating to “get stressed out easily” unconsciously self-select towards an environment that suits this trait: entrepreneurship. Furthermore it is remarkable that the Summary Statistics in Table A2 show that the standardized Emotional Stability mean is higher for the incorporated business owners than for the salaried workers, whereas the coefficient of Emotional Stability is significantly negative in our regression in Table A3. We will leave this observation open to investigation by future research due to the exploratory scope of the research at hand.
Significant evidence for hypothesis 2c; Hypothesis 2b is rejected. Our results on Conscientiousness confirm only one hypothesis and reject the other. Significant at the
10% level in our comprehensive model, a higher Conscientiousness seems to decrease the probability of becoming an entrepreneur. Where this trait was in line with the popular perception of an entrepreneur’s need to achieve, work hard and be persistent (Zhao et al., 2010), the effect may be overshadowed by a decrease in both her flexibility to react to crises and her freedom to operate and experiment. Therefore, our results are in line with Leibenstein (1968)’s gap- and input filling role and the way Schumpeter considers the role of an entrepreneur: providing creative leadership. Further research may look into the division of tasks within incorporated businesses in the Netherlands and see what tasks the owner is concerned with as compared to their employees. It may be reasonable to argue that those tasks requiring a high degree of Conscientiousness can be delegated to employees so that the entrepreneur can flexibly solve problems or lead the firm in a new direction. Secondly, according to the person- environment fit theory it may be the case that those persons scoring high on the conscientiousness may self-select into salaried employment where their tasks are well-defined, and that those low on the scale are looking for the freedom that self-employment offers. Extra evidence for this explanation might be the high percentage of entrepreneurs who express that the independence and freedom associated “being one’s own boss” has cause them to select this form of occupation (for more information see Begley & Boyd, 1988; Gatewood, Shaver, & Gartner, 1995; Turan & Kara, 2007).
No significant evidence for hypotheses 2a, 2f and 2g. Our results do not provide
conclusive evidence about the effects of Agreeableness and Extraversion on
entrepreneurship. Nevertheless, the direction of both effects is in line with what our hypotheses and literature study suggested. And when not controlling for family
background, both factors become significant. Extraversion may be less important than anticipated because those who start a new company likely need to be strong on the content and not necessarily on the people skills. Future research on the Netherlands may focus on the difference between founders and managers in this aspect18, or try to link the need for Extraversion to larger firms where the need for self-expression and leadership is higher (Judge & Bono, 2000).
Significant evidence for Hypothesis 2h. The Rosenberg Self-Esteem indicator has a
relatively large coefficient (p=0.095) in our comprehensive model. This relationship is in line with our hypothesis and literature in this field and will probably become more significant in a larger sample. A second reason for a larger standard error than in other empirical studies may be the high correlation with Emotional Stability.
Hypotheses 2i is rejected, no evidence for Hypothesis 2g. We decided to include the
Need for Cognition measure, a variable that is not regularly used in literature about entrepreneurship. The effect is not significant and goes in the opposite direction than
18Brandstätter (2011) compares the big-five personality traits of managers to those of entrepreneurs; A study that could be replicated in the Netherlands.
expected. This might be due to its relatively high overlap with what Openness to Experience measures (r=0.65). Also, there is little empirical evidence for its stability across a life span, and the fit of the model does not improve when including Need for Cognition (compare Table A3 to Table A4). We can therefore conclude that it is not worth including into our model of entrepreneurial intentions and its general usefulness in this context is questionable.
No significant evidence for Hypotheses 3a and 3b. Our results about family
background provide some surprising results, namely that Father’s education has a significantly positive effect on the likelihood of becoming an entrepreneur and the mother’s education has a negative effect, although not significant. Being contradictory to the situation in for example the United States (Levine & Rubinstein, 2013) this effect of maternal education in the Netherlands may be worth investigating further as it might be a cultural feature of the Dutch family situation.
Implications regarding classical theories. The above conclusions also allow us to discuss the validity of the classical theories and the entrepreneurial roles. Some characteristics of entrepreneurs are supported by all theories and backed by significant coefficients, e.g. Self-Esteem and Openness to Experience. But other conflicting
hypotheses allow further interpretation. Firstly, there is evidence to reject the notion that the entrepreneur is simply a capable coordinator of production. Low scores on
Conscientiousness and Emotional Stability actually have a positive effect on the
likelihood to become an entrepreneur. This could mean that flexibility and creativity are more important than responsibility towards duties; speaking in favour of Schumpeter (1939)’s creative leadership role, and opposing Say (1836)’s role as the end-responsible
production coordinator. Secondly, Extraversion and Agreeableness have no significant effect, which indicates that some effects proposed by the theory might not hold for the whole entrepreneurial base. Thirdly, it may not be the Emotional Stability that permits an individual to become entrepreneur, but rather the opposite Neuroticism that attracts individuals towards this uncertain occupation. This stands contrary to most classical theories that describe entrepreneurs as some form of a “superior kind of labor” only exceptional individuals are good enough to engage in (Say, 1836, p. 330). Yet, our results in combination with the Person-Environment Fit theory suggest that neurotic individuals are simply more drawn towards uncertainty. In summary, the results are more in line with Schumpeter (1939)’s role of the entrepreneur as the creative leader of a firm than with the notion of the entrepreneur as a capable and responsible coordinator of
production (as proposed by Say, 1836). Speaking more broadly, evidence suggests that for the entrepreneurial role flexibility is more important than dependability.
Limitations and further research
The results presented in Table A3 show evidence for a number of our hypotheses, for which we will discuss possible validity concerns. Subsequently, we will turn towards open questions. As this is an exploratory study under the limitation of secondary data, we will continue by formulating our recommendations for topics of future research and recommend measures to improve the quality of data collection. In the section afterwards, we will conclude.
A first limitation is that the sample does not allow us to isolate entrepreneurial intention from entrepreneurial success. Measuring success is not possible as there are too few details available about the firm19. Neither size, growth, nor profitability are given in the dataset: The only possible measurement is the fiscal profit (or loss), which is too limited as a measurement and prone to underreporting and timing-related biases. Our sample therefore might be biased towards unsuccessful new entrants that fail at some point with their venture. Upon an investigation of the number of entrepreneurs that quit (meaning they leave incorporated self-employment again), we see that only 68 out of 221 or 26,24 % of entrepreneurs fall have abandoned their incorporated business again, close to the Dutch failure rate of 6 % per year that the European Commission published in their report (The-Gallup-Organization, 2008). Even though we do not distinguish between entrepreneurial intentions and success, this finding makes it safe to interpret the results as speaking for a representative sample of entrepreneurs in the Netherlands.
A second valid consideration is that some scholars suggest that entrepreneurs are too heterogeneous a group for our analysis to produce any reliable results. Gartner (1988) states that differences among entrepreneurs and among their ventures can be as great as the variation between entrepreneurs and non-entrepreneurs. Differences can be found between for example industry, size, and age of the venture. In anticipation of this
problem we focus on incorporated businesses only, which reduces variation substantially. Regardless, some of the results are likely explained by differences between businesses e.g. in the case of Extraversion. A second distinction is between entrepreneurs and managers of a firm, as brought up by Amit et al. (1993). We have relied on multiple theories of entrepreneurship and selected incorporated business owners, not hired managers, as the group of interest for our analysis. Future research could focus on the differences between managers and owners of businesses in a Dutch sample.
A third note-worthy point lies in the large number of variables included in the study, which increases the likelihood of finding type I errors. Moreover, an increase of type II errors is also possible due to multi-collinearity between different variables of the regression model. In response, we also present the results of three models that include
19(Murphy, Trailer, & Hill, 1996) have dedicated an article to the measurement of performance in entrepreneurship research. They argue that there is a large number of possible measurements and variables to capture them: efficiency, growth, profit, size liquidity, market share, to name a few.
only a subset of the variables to show the validity of their constructs separately. However, as strong support was found for many variables in the comprehensive model, we assume that this limitation did not distort our results.
A fourth important limitation is the threat of endogeneity within our model, which can affect the interpretation of our results in two ways. Firstly, it might be that actually being a business owner influences our assumed-to-be independent variables. This reverse causality may distort the interpretation of the aforementioned results. Though we have selected the variables accordingly20, some personality measures may change from
entrepreneurship experiences. However, the Big Five personality measures are found to be relatively stable over a lifetime (Cobb-Clark & Schurer, 2012). This would allow us to draw conclusions that those born with the stipulated traits are more likely to self-select into entrepreneurship, and not that being an entrepreneur makes people for example more open to experience. A similar level of stability has been reported for the Rosenberg Self-Esteem score, although not as conclusive as for the Big Five. Self-Esteem can alter during the relatively instable years of adolescence, whereas afterwards it is shown to stabilize similar to other personality traits (Trzesniewski, Donnellan, & Robins, 2003). Yet our age distribution in Figure 1 shows that most entrepreneurs engage in
entrepreneurship after adolescence when personality traits have mostly stabilized. Little is known about the stability of the Need for Cognition, another drawback of this
measurement. In summary, we can state that reverse causality likely does not pose a large enough risk to draw our results in dubio.
Another cause of endogeneity is that an unobserved third variable influences both the tendency to become an entrepreneur and an assumedly independent variable, as mentioned in our theoretical framework under challenges. For example, the memoirs of famous entrepreneurs often explain their unique business attitudes through childhood experiences or early exposure to business and technology (as described by for example Branson, 2007; Walton, 1993). These experiences might have proven formative for both the entrepreneurial spirit and the personality and thereby lead to an observed correlation. While the correlations are open for interpretation and discussion, we cannot infer with certainty that there must be a causal relationship observed.
Open Questions. This explorative study of the characteristics of Dutch entrepreneurs leaves a number of open questions that demand further investigation. Firstly, an enquiry into country differences between the Netherlands and other countries may give interesting results. The importance of spatial differences has often been
discussed, and Short et al. (2010) have described recent advances in methodology that make this analysis promising for future research. The European Commission
(The-Gallup-Organization, 2008) and the Global Entrepreneurship Monitor (van der 20High personal wealth, for example, may both facilitate entrepreneurship, or be the result from it. We therefore instead included family wealth into our analysis, which can normally be considered exogenous.
