HOW ENTREPRENEURIAL CHARACTERISTICS INFLUENCE COMPANY CREATION: A CROSS-NATIONAL STUDY OF 22 COUNTRIESTESTED WITH PANEL DATA METHODOLOGY

This study analyzes, from a multicountry perspective, the in uence of the sociodemographic pro les of nascent and new entrepreneurs on their behavior. The panel data-based research approach combines temporal series and cross-sectional data to assess entrepreneurial activities across 22 countries with varying income levels. The results show that entrepreneurs’ characteristics in uence entrepreneurial behavior signi cantly and positively, in the following order: previous experience of the founder, age, and education. These ndings suggest valid recommendations for stimulating entrepreneurship, both for enterprising business founders and for the institutions responsible for designing economic and regional development policies.


Introduction
Research into entrepreneurs is critical for understanding the creation of enterprises (e.g., Frederking 2004;Grüner 2006;Shirokova, Knatko 2008;Uhlaner, Thurik 2007;Westhead et al. 2005), especially in contexts marked by a large number of small and medium-sized enterprises (SMEs). In these scenarios, the type of employment often determines the evolution of the productive unit (Grüner 2006;Johnson et al. 2006;Nikolaus, Christian 2004;Van Praag, Versloot 2007), and therefore, extensive research focuses on assessing the gure of the entrepreneur (e.g., Arenius, Minniti 2005;De Jorge Moreno, Laborda Castillo, De Zuani Masere 2010;Shirokova, Knatko 2008;Wagner 2007), including how the entrepreneur's pro le affects the recognition and exploitations of business opportunities (e.g., Zhao et al. 2009). This line of research encompassess different theoretical approaches (e.g., economic, sociological, psychological, managerial) but still is not exempt from certain disadvantages, especially pertaining to the methodology and type of sample used to study the associated phenomena (Reynolds 1997;Van Praag, Versloot 2007;Volkman 2004). Most prior work considers relatively small samples from a single country, without noting their gender or socioeconomic environment. Yet the entrepreneur's attitudes, approaches, values, and satisfaction with the status quo have remarkable impacts on the creation of the company (e.g., Arenius, Minniti 2005;Frank et al. 2007;Frederking 2004;Furkukawa et al. 2007;McCline et al. 2000;Uhlaner, Thurik 2007).
Moreover, prior research has not clearly de ned the various stages of entrepreneurship, combining nascent, new, and consolidated entrepreneurs within the same category, which may generate biases in the results. The lack of distinction between the previous stages of the entrepeneurial activity likely creates bias, and previous research (e.g., Reynolds et al. 2005) notes the need to consider different types of entrepreneurs according to the time passed since the creation of the business. First, nascent entrepreneurs have completed the rst process of entrepreneurship, moving from the conception of the business into the gestation or start-up process. They generally have paid salaries and wages for less than three months. Second, new entrepreneurs have been in business (i.e., paid salaries and wages) for at least three months but less than 3.5 years. Third, established entrepreneurs have been in business for a while, that is, at least 3.5 years.
This study therefore identi es the sociodemographic pro les of entrepreneur to determine how they affect enterpreneurial behavior. Furthermore, it employs and extends the scheme proposed by Arenius and Minniti (2005) and incorporates contextual variables related to the environment. In particular, this study includes nascent and new entrepreneurs, that is, those whose companies' duration does not exceed 42 months. The sample of entrepreneurs spans 22 diverse countries, located on four continents, and therefore can identify cross-national differences according to the gender of the entrepreneurs and their income levels (measured as each country's gross domestic product [GDP]). Finally, to analyze these data, this study uses a panel methodology, which provides added interest by combining temporal series and cross-sectional data.
This article consists of ve further sections. The next section reviews theory related to entrepreneurship and related hypotheses for this investigation. After a description of the methodology, the following section details and discusses the main results. Finally, the conclusion synthesizes the principal implications and proposes a series of recommendations for both entrepreneurs and the agents in charge of developing economic policies.

The in uence of the entrepreneur's age
Entrepreneurship literature highlights the important impact of the age of the entrepreneur with regard to his or her experience and vitality (De Jorge Moreno et al. 2010;Sandberg, Hofer 1987). In general, older people likely enjoy an advantage derived from accumulated experience, such that the likelihood of creating a new company increases among older entrepreneurs (Hesselset et al. 2008;Levesque, Minniti 2006). Yet other research shows that younger people tend to possess more energy, vigor, and enthusiasm, which in combination with knowledge and open mental attitudes enables them to seek, recognize, and develop more innovative business opportunities with economic growth potential (Grilo, Irigoyen 2006;Hessels et al. 2008). Therefore, this study posits that over the course of time, a person accumulates experience but also loses aptitude for recognizing opportunities (e.g., possesses less updated knowledge, prefers to avoid risk, suffers personal and psychological limitations). In turn, this study offers a nondirectional hypothesis regarding the in uence of the age of an entrepreneur:

H1:
The age of the entrepreneur in uences in the creation of a business.

The in uence of the entrepreneur's education and experience
Education also appears associated with entrepreneurship, especially with regard to the bene ts of education for the level of managerial action and the pro tability of the company, among other indicators (Arenius, De Clercq 2005;De Jorge Moreno et al. 2010;Naude et al. 2008;Sapienza, Grimm 1997). Education or training generally is associated with the development of capacities (e.g., analytical skills, information processing aptitude, idea association) that contribute to the recognition and development of innovative business opportunities. People with an advanced degree are more inclined to introduce innovations into the market (e.g., Klandt 2004), develop intentions to create their own companies (e.g., Oosterbeek et al. 2009;Peterman, Kennedy 2003;Raguseo 2009;Župerka 2010), and detect business opportunities with innovative potential (e.g., Arenius, De Clercq 2005;Souitaris et al. 2007).
Because, as some authors indicate, entrepreneurial success depends on both the entrepreneur's personal characteristics and professional aptitude acquired through education (Naude et al. 2008), most entrepreneurs possess a high level of education. Roberts (1970) nds that the founders of high-tech companies tend to possess at least a master's degree, and Cooper and Dunkelberg (1987) note that entrepreneurs' education level is higher than that of the population in general. Gelderen et al. (2001) and Naude et al. (2008) also suggest that most entrepreneurs possess a high education level.
Noting these concepts, and considering the expansion of education in recent years, which has demonstrated the positive effects of higher education on the ability to start a company, especially in competitive environments, we propose:

H2:
The younger the entrepreneur, the greater his or her consciousness of the bene ts for higher education for creating a new business.
Some ndings reveal that a founder's education, experience, and previous knowledge relate positively to both the emergence of companies and managerial consolidation (Davidsson 2006;Hyytinen, Ilmakunnas 2007;Lee, Tsang 2001;Naude et al. 2008;Sapienza, Grimm 1997;Stam et al. 2008;Ucbasaran, Westhead 2002). Speci cally, businesspeople with experience (both managerial and labor) should recognize more innovative business opportunities (Ucbasaran, Westhead 2002) and be more inclined to undertake related work activities (Cooper, Dunkelberg 1987;Kolvereid, Isaken 2006;Stam et al. 2008), often in response to an interaction with the environment that generates an impulse to explore an idea (Alas 2004;Zander 2004). For example, preexisting and ongoing relations with important clients might provide valuable opportunities for the entrepreneur to join a commercial network and attract loyal clients in the long term (Ripollés, Blesa 2006;Shrader et al. 2000). In this case, experience provides a valuable human capital resource for company creation and consolidation (Peter, Vertinsky 2008). Therefore,

H3:
The degree of previous experience of the entrepreneur (entrepreneurial, technical, or managerial) positively in uences business startup.

Methodology
To con rm the proposed hypotheses empirically, this study employs the Global Entrepreneurship Monitor (GEM) database and gathers a sample of 22 countries. This database and the associated project enjoy a strong reputation, especially for entrepreneurship studies, and provides nearly innumerable possibilities for obtaining information partitioned and differentiated according to diverse variables.
This study focuses on the period between 2002 and 2006, during which time all 22 sampled countries took part constantly in the GEM project. According to the GEM classi cation by income (i.e., levels of GDP per capita), this sample includes high, medium, and low income countries, as follows: • High income: United States, Canada, United Kingdom, Denmark, Finland, France, Belgium, Germany, Ireland, Italy, Japan, The Netherlands, Norway, Singapore, Slovenia, Spain, Sweden. • Medium and low income: Argentina, Brazil, China, Croatia, South Africa.
The variables in the empirical model re ect information obtained in 589,377 surveys of persons between the ages of 18 and 64 years who declared their involvement in an entrepreneurial project. Table 1 outlines these variables. The proposed general model includes panel data and xed effects: This model also can be disaggregated into separate models for men and women to discern any differences related to this factor. Table 2 contains the variables used in these models.
The equations for these gender-speci c models are as follows:  (2) The education of primary level offers to the children basic competitions of reading, writing and mathematics, besides an elementary comprehension of topics as history, geography, natural sciences, social sciences, plastic arts and music. The education of secondary level completes the presentation of basic education that began in the primary level and points to lay the foundations for a learning and a human development that they last the whole life, offering an instruction more orientated to topics or speci c competitions, with more specialized teachers. The tertiary, medium education or post secondary, though it does not represent the step to a quali cation of advanced investigation, normally demands, as minimal requirement of admission, the satisfactory ending of the education of secondary level. The top education, where the sciences and the investigation would have major content, case of the diplomaturas, masters, masters, conferred a doctor's degree and post doctorates. Source: Institute of Statistics of the United Nations Organization for the Education, the Science and the Culture (UNESCO). Panel data methodology, which combines temporal series and cross-sectional data, is appropriate for estimating the behavior of nascent and new entrepreneurs across countries, because it can offer estimations even when no observable heterogeneities emerge for each country or across time. In addition, it can identify individual-speci c effects for each entrepreneur pertaining to how he or she decides to create a company.

Analysis of results and discussion
The estimations of Equations 1-3 rely on estimated generalized least squares (EGLS), corrected for heteroscedasticity by White's method using the variables in the general model (see Table 3). The variance explained by the endogenous variables TEAit, TEA-MALit, and TEAFEMit are 90.75%, 83.9%, and 88.3% respectively.  Table 4.
These elasticities reveal the order of importance of the characteristics of nascent and new entrepreneurs across 22 countries: previous experience of the entrepreneur, with an elasticity of 0.4818%; age of the entrepreneur, especially for entrepreneurs between 18 and 34 years of age, with an elasticity of 0.2534% and for those between 55 and 64 years of age, with an elasticity of 0.0763%; and nally, a medium level of education, with an elasticity of 0.0175%.Yet some differences emerge for men and women. The average male entrepreneur has experience (elasticity = 0.15%) and is aged either 25-34 years (elasticity = 0.098%) or 55-64 years (elasticity = 0.097%); he also has earned secondary and medium education levels (elasticity = 0.0897%). Table 3 A. Álvarez-Herranz et al. How entrepreneurial characteristics in uence company creation ...  For female entrepreneurs, experience again has the greatest effect, with a similar elasticity (0.16%). However, the next most relevant characteristic is education, especially medium and superior levels (elasticity = 0.08%). Age also has an effect, such that the age interval from 35 to 44 years indicates an elasticity of 0.11% (elasticity of the variable TEAF3544), followed by 18 to 24 years, with an elasticity of 0.0689% (elasticity of the variable TEA1824), and then 25 to 34 years, with an elasticity of 0.0422% (elasticity of the variable TEA2534). Therefore, these ndings suggest that entrepreneurial women, on average, decide to start business when they are between 18 and 44 years of age, which is opposite the results for men, who initiate their businesses either between 25 and 34 years or between 55 and 64 years of age.

End of
To detect further differences in the models that may re ect country characteristics, the next model estimation classi es the sample according to the income level of the country. The results are in Table 5, and the associated elasticities for each of the variables, according to gender and income level of countries, are in Table 6.
The analyses in Tables 5 and 6 reveal that entrepreneurs in high-income countries, in average terms, have experience (elasticity = 0.556%) and range in age from 18 to 34 years (elasticity = 0.204%) or 55 to 64 years (elasticity = 0.0995%), with medium and superior educational levels (elasticity = 0.105%). In medium-and low-income countries, entrepreneurs again represent two main age groups with similar elasticities, namely, 55-64 years (elasticity = 0.199%) and 18-24 years (elasticity = is 0.196%). The characteristic that differentiates these two groups is experience; it is signi cant for the 55-64 year age group, with an elasticity of 0.545% (variable TEASUB), but not for the 18-24 year age group. That is, an older entrepreneur has more experience and initiates a new venture in response to an opportunity rather than need. The ndings reverse for the youngest group of entrepreneurs, who lack experience and consequently create companies due to their need-based motives. Although the level of education of the average entrepreneur in medium-and low-income countries is secondary or average, the latter predominates over former level, as re ected in the negative elasticity (-0.153% and -0.142%, pertaining to the variable TEASD) of an medium education level. Thus, as these countries improve their educational level, the number of entrepreneurs with a secondary level of education diminishes and the proportion with an average level increases, especially among young entrepreneurs. Accordingly, the elasticity of this variable TEAPS is high (0.196%) for the group of 18-24 year olds compared with that among the group of 55-64 years old (0.155%).
As for the men's educational level, the secondary-medium education level stands out (elasticity = 0.1147%), whereas female entrepreneurs exhibit a higher level of studies (elasticity = 0.1532%). This indicator may reveal that female entrepreneurs start their businesses later than men, apparently because they have obtained suf cient education. Finally, experience is a determinant factor for both men and women making the decision to begin a business. However, the variable has more weight for men than women, according to their respective elasticities (0.3121% versus 0.0750%).
These characteristics also differ for high-versus medium-and low-income countries.
In the latter group of countries, the average male entrepreneur is younger (between (1) The third column under the general model features entrepreneurs aged between 55 and 64 years, for which the relevance of experience was hidden in the general model by the higher weight in the sample of entrepreneurs aged between 18 and 24 years 18 and 24 years, elasticity = 0.6305%), but older male entrepreneurs (55-64 years) in high-income countries show a higher elasticity (elasticity = 0.2622%). Women in both groups of countries reveal a signi cant elasticity only for the 18-24 year age group, though the elasticity is higher in medium-and low-income compared with high-income countries (0.1374%).
Finally, as for the educational level, in this group of countries, both male and female entrepreneurs possess an average educational level. The more in uential factor for women is experience, whereas men indicate a negative elasticity for the experience variable (elasticity of -0.5196%). Men in these countries start businesses due to their need, not in response to opportunity, such that they sense an obligation to start a business despite lacking experience. The opposite effect occurs for women. Table 7 synthesizes the hypothesis results. These results indicate that female entrepreneurs start their businesses later than men, apparently because at that point, she has obtained suf cient education. Yet the in uence of experience is not that different on the behaviors of men versus women. Both male and female entrepreneurs recognize the importance of previous experience for the success of their start-up project and work to detect and develop new business opportunities with innovative potential. Previous experience clearly helps potential entrepreneurs nd innovative business opportunities and develop a network of contacts that can provide advantages for the creation and consolidation of their company. Such experience also may provide the entrepreneur with the con dence needed to initiate entrepreneurial activities.
This study also notes the effect of the level of country income, such that the characteristics of average entrepreneurs vary somewhat. The most relevant difference between high-and medium-and low-income countries for men relates to experience; older entrepreneurs have more experience than the youngest entrepreneurs, and older male entrepreneurs start their companies in sectors they know, so they can take advantage of business opportunities based on their previous experience. These results suggest though that regardless of the economic level of the country, a growing number of older workers resist de nitive retirement and regard entrepreneurship as a means to remain active in the labor market. For women, country income reveals a greater age difference: In high income countries, female entrepreneurs are mostly 35-44 years, whereas in mediumand low-income countries, entrepreneurial women tend to be 18-24 years of age.
This model incorporates diverse variables related to the sociodemographic pro les of nascent and new entrepreneurs in 22 countries across four different continents and analyzes them according to both gender and the income per capita in their country. It therefore represents a key contribution to entrepreneurship research. Moreover, this study reveals the importance of sociodemographic variables in stimulating the creation and consolidation of companies.
Broadly, this research offers several contributions. First, in contrast with most prior investigations, which analyze entrepreneurs using samples from only one culture, this study considers a very wide sample from 22 countries that ensures great cultural diversity. Second, this study identi es which sociodemographic characteristics of men and women at different socioeconomic levels have major impacts on their entrepreneurial behavior. Therefore, government agencies dedicated to increasing entrepreneurship rates in their countries can better determine which groups to focus on in their communication and recruitment efforts. Third, this investigation focuses on a very concrete phase of entrepreneurship, unlike previous work that fails to distinguish nascent, new, and consolidated entrepreneurs. By limiting the sample to nascent and new entrepreneurs whose companies have existed for less than 42 months, this study diminishes possible heuristic biases derived from the joint consideration of different phases in the creation of a company.
However, this study also suffers several limitations. In particular, the constructs in the causal model are not comprehensive; they constitute only a portion of the diverse prec-edents and consequences that could have been considered. However, to avoid excessive model complexity, these factors, precursors, and mediating variables have been limited purposefully. Further research should incorporate additional new variables in the model (e.g., public help programs, social networks, cultural values, business expectations). It also would be interesting to examine the type of planning that young versus older entrepreneurs undergo, the degree of innovation in the opportunities they detect, and the motivations that primarily stimulate them to create companies. For example, exactly what motivates men versus women of 55 to 64 years of age to create a company?