The Entrepreneurial life course of men and women Karl Wennberg Abstract: This paper presents a perspective on entrepreneurship that integrates life course theory and capital theory to derive predictions for how human, social and cultural capital interacts with life course events, shaping the likelihood of entrepreneurial entry and exit. The predictions are tested on a random sample whose life courses are reconstructed using archival data. Results provide strong support for our model and suggest a more dynamic view of the entrepreneurial process than is offered by capital theory. Our model challenges the view of entrepreneurship as an occupational choice driven by the rational decisions of individuals to exploit their human capital. Center for Entrepreneurship and Business Creation Stockholm School of Economics Box 6501 SE-113 83 Stockholm, SWEDEN Phone: +46 8 736 9356 Email: [email protected] Introduction A wide range of traditions in the social sciences investigate entrepreneurship as a career choice. Cursory explanations from a

psychologically oriented perspective see

entrepreneurial career choices as “planned” or “intentional” behaviour based on individual motivations, psychological make-up (Hao, Seibert & Hills, 2005), and career aspirations (c.f. Krueger, Reilly & Carsrud, 2000).Economic and sociological models of rational choice, on the other hand, see entrepreneurship as an occupational choice that maximizes the individual‟s social or economic utility (Evans & Jovanovic, 1989), while sociological models of status-attainment highlight the role of intergenerational social and cultural capital (Sørensen, 2007). These and other career choice models have informed research on a variety of individual-level, family-level and external factors associated with entrepreneurship. However, the predictions from these models are often tested in cross-sectional analysis, based on entrepreneurs sampled before they made this career choice (Cornelius, Landstrom &

1

Persson, 2006). Little work has been done on why and how individual entrepreneurial events unfolds under an individual‟s life course, or why this is so. I know from prior research that the processes of entrepreneurial entry and exit are often affected by life course events, such as ageing, geographic moves, switching careers and partners: in Aldrich and Kenworthy‟s (1994) words “Entrepreneurship happens when people are on their way to something else”. In this paper I contend that entrepreneurship is most often not a conscious career choice, nor is it particularly planned and frequently is not mainly intentional. The social processes and the societal positions in which individuals find themselves following life course events increase the likelihood that advantageously placed individuals will discover opportunities, sometimes through active search and sometimes simply through their being in the right place at the right time, and with the right stock of knowledge (Baker et al., 2003). I outline a theoretical model based on life course theory and capital theory. Life course theory frequently is applied in sociology to investigate the patterns and ordering of individual events, such as marriage, geographic relocation, occupational changes and childrearing, but has received scant attention in the entrepreneurship literature. 1 Capital theory predicts that entrepreneurial opportunities are created at the intersection between individual agents and their social connections, where individual career choices are formed by their education, training and cultural background. Yet, these theories generally do not take account of how life course events affect the probabilities of opportunities being actively exploited. I integrate these two theoretical perspectives to derive predictions about how the accumulation of human, social and cultural capital interacts with life course events, such as marriage, geographic moves, occupational changes, and childrearing, to affect probability of engaging in, or abandoning, entrepreneurship. These predictions are tested on a random

1

But see Carroll and Mosakowski (1987) for an important exception.

2

population of individuals whose life courses are reconstructed using archival data. Event history models reveal that while human and cultural capital matter greatly for the entrepreneurial processes of entering and exiting new ventures, such capital resources are highly contingent on the life course events of marriage, childrearing and geographic mobility. This research contributes to the entrepreneurship research in three ways: empirical, conceptual, and methodological. First, I contribute empirically to research on the entrepreneurial process as being moulded by the accumulation of capital, by integrating these more “static” models with more serendipitous life course events. Our model challenges the view of entrepreneurship as an occupational choice driven by the rational decision of the individual to exploit his or her human capital (e.g. Douglas & Shepherd, 2000). Second, our model contributes conceptually in highlighting a dynamic view of the entrepreneurial processes as being shaped at the intersection between the resourceful individual and opportunities enacted by life course events. The theoretical model do not assume the exploitation of such opportunities as being neither planned nor optimized, but when wilfully pursued leads the individual to create a new firm, potentially spreading their ideas and innovations. Third, in using a longitudinal research design based on archival data, our research helps to mitigate a number of common biases in the field (Davidsson; 2004 Carroll & Mosakowski, 1987; Shane, 2008). Theory and Hypotheses Capital theory, originating partly in sociology and partly in economics, holds that the accumulation of human, social and cultural capital leads to the build-up of a “stock” of personal resources that can be leveraged by individuals to achieve entrepreneurship in the shape, for example, of starting or growing a new venture (Kim, Aldrich & Keister, 2006). This strand of research often compares the impact of various capital resources on career choices, in an attempt to identify which types of capital matter most for particular choices. 3

For example, the study by Kim and colleagues (2006) finds that human capital is a stronger determinant than financial or cultural capital resources for entrepreneurial entry. Other research shows that human capital can be a substitute for financial capital (Bates, 1990), and that social capital can be a substitute for human capital (Davidsson & Honig, 2003). In this paper I use the sociologically-flavoured capital theory inspired by Bourdieu (1977), which holds that the role of capital resources lends itself to the social structuration of work, whereby human capital is based on education and training, and cultural capital is gained by the intergenerational inheritance of norms and skills. The cultural capital provided by parents shape entrepreneurial career opportunities both through exposure to entrepreneurship as a viable career choice, and through social closure, meaning that children are expected to follow in their parents‟ footsteps, perhaps by taking on the family business (Sørensen. 2007). Hence, our first set of hypotheses relates to the intergenerational transmission of cultural capital, and the potentially contingent forces emanating from life course events. First, cultural capital is transmitted from parents to children in terms of the latter adopting a similar career path to that of their parents (e.g. Kim et al., 2006; Sørensen. 2007), suggesting that: H1a: Parents’ entrepreneurship will have a positive effect on entry into entrepreneurship.

However, the general process of intergenerational transmission of cultural capital is often posited as differentially structured across social and ethnic strata (Steinmetz & Wright, 1989). Giddens‟s (1984) theory of structuration holds that individual agents and social systems co-evolve and that individual agents are simultaneously enabled and constrained by the social structuring of their environment. Perhaps the strongest source of social support and social constraints posited in sociological research is that of upbringing and social class. This line of research argues that societal structures flow both vertically along generations, and

4

horizontally along social classes, such that the way that cultural and social capital is transmitted from parents to children is contingent on the social and ethnic stratification of the society. Butler and Herring‟s (1991) study support this view: they investigated 7,542 persons aged over 18 from the US General Social Surveys 1983-1987 and found that among those with Jewish ancestry, father‟s self-employment was positively related to children‟s selfemployment, but among those with Polish or African ancestry, it was negatively related to children‟s self-employment. This suggests that capital resources may not be universally salient, but rather that the inheritance of cultural capital is contingent on the structuration of society across ethnic lines. I therefore hypothesize: H1b: The positive effect of parents’ entrepreneurship will be stronger for individuals from a minority cultural background.

Life course dynamics and the entrepreneurial processes Theoretical work in life course dynamics holds that lives are influenced by their changing social contexts, and that it is these contexts that are important to study how dynamic patterns evolve over time (Blossfeld, 1986). Here, “life course” is conceptualized as the path and the crucial elements that form the multiple trajectories of individuals and their implications in terms of socio-economic development (Elder, 1998). Research in this vein focuses on a variety of issues such as how family-related factors affect labour market behaviour and how people react to life events (Schroots, 2003). Central to this research tradition is the emphasis on age, individual agency and intergenerational processes, for how lives evolve (Elder, 1994). Some more recent life course research emphasizes “turning points” in explaining behavioural change over the life course (Uggen, 2000). This process, which evolves along the education and career path for example, is “interrupted” by significant life course events related perhaps to family, health, etc. In entrepreneurship, life course dynamics 5

entails investigation of the process through which entrepreneurial processes are shaped, or influenced, by the life course events during the different stages of their lives and careers (Carroll & Mosakowski, 1987). This research has produced insights into how life course events interact with social background in moulding careers and life paths (e.g. Elder, 1985). Life course dynamics and gender in the entrepreneurial processes Among the most important dynamics affecting the life course are marriage and cohabitation. In historical societies, integrating one‟s life and career path with that of another person was often an economic necessity, or was the means to join two families rather than being the outcome of an emotional union between two persons (Padgett & Ansell, 1992). In modern societies, emotional bonds lead to marriage or cohabitation and the integration of lives and careers, sometimes involving implicit bartering or trade-off between household work and career between the partners, or facilitating of each others‟ career prospects by providing the resources, information and social support for career choices. This suggest that life course theories are related to the basic capital resources that individuals gain through their upbringing, and accumulate over time, in that life course events represent contingencies that may propel or inhibit career choices such as those related to entrepreneurship. In the particular context of marriage and cohabitation, gender theory plays a role in explaining how gender roles are produced, reproduced, and integrated in individual life courses through encouragement of encouraging socially accepted and gender-normative types of behaviour, and discouraging behaviour that is considered less appropriate to a given gender (Papalia & Olds, 1981). For the purpose of the model of entrepreneurial life course, gender roles reflect the implicit division of labour that exists in contemporary society, by operating as mediators for how human and social capital are mobilized into economic action. I therefore predict that: H2a: Marrying/becoming cohabitant a man with current or prior entrepreneurial experience will have a positive effect women’s probability of entry into entrepreneurship. 6

H2b: Marrying/becoming cohabitant a woman with current or prior entrepreneurial experience will have no impact on men’s probability of entry into entrepreneurship.

One of the major differences in the gender roles, especially as they impact on labour markets, is childbearing and rearing. There is a long tradition of studies on the impact of children on career choices, in economic and sociological work. For our interest in the entrepreneurial life course, I draw specifically on the research investigating how the arrival of children into the household affects the propensity to engage in entrepreneurship among household heads. Prior research, such as Boden‟s (1999) study of 37,000 people from the US Population Survey 1987-1991 finds women‟s engagement in entrepreneurship positively associated with the number of children. Folta, Delmar and Wennberg (2010) studied the labour market dynamics for individuals engaged in part-time and full-time entrepreneurial entry. In a sample of 44,033 Swedish men followed between 1994 and 2001, they found that the number of children was positively associated with both types of entrepreneurial entry. Based on this tentative evidence and the prediction of our theoretical model that life course events such as children represent strong contingent forces that shape how human and cultural capital are mobilized in start-up efforts, I posit the following hypotheses: H3a&H3b: Having children will positively influence men’s & women’s probability of entry into entrepreneurship.

The role of gender is evident in entrepreneurial choices in terms of how the arrival of children often leads the household head to reconsider his or her social and economic career goal and in the household‟s internal division of labour, which may affect the propensity to engage in entrepreneurial activity, but also may shape the role structure of extant entrepreneurs. The arrival of children often means that the breadwinner in the family will experience changes in their perceptions of the rewards and responsibilities that accompany 7

their work. For example, Gimeno and colleagues (1997) found that being a parent affected the performance thresholds of these entrepreneurs in that it affected the perceived “switching cost” of exiting from entrepreneurship. Ironically, this suggests that in societies where the male is the breadwinner, still true for most modern economies, male entrepreneurs might have lower performance thresholds in that they feel more committed to their role of entrepreneur. I therefore hypothesize that: H3c: Controlling for entrepreneurial earnings, having children will negatively influence men’s probability of exiting from entrepreneurship.

The arrival of children in the household is known to impact differently on men‟s and women‟s career moves and career prospects. Among women it has been shown that childrearing career women report more “role anxiety” and stress after becoming a parent (Wiley, 1991). There is a body of research that suggests that self-employed women are disadvantaged due to the barriers related to social norms and role expectations, the division of labour in families, and the level of social support (Aldrich, 1989; Brush, 1992). Boden and Nucci (2000) surveyed 4,881 entrepreneurs from the 1982 and 1987 Characteristics of Business Owners in the US. They found that the higher exit rate of women largely depended on their smaller availability of financial capital to start their business, but did not measure the presence of children in the household. The evidence from Glass and Fujimoto‟s (1995) study of pregnant entrepreneurs in Indiana, suggests that the ability to work flexibly from home is a major factor in women‟s decisions to remain in entrepreneurship. This suggests that the parental role structures of individuals already engaged in entrepreneurship are differentially affected by the arrival of children in the household. I hypothesize therefore that: H3d: Controlling for entrepreneurial earnings, having children will positively influence women’s probability of exiting from entrepreneurship. 8

Geographic moves and the entrepreneurial life course Hitherto, our theoretical arguments have focused primarily on intra-family events. In terms of extra-family events, I know that geographic mobility is an essential component of life course dynamics. Out-of-area moves entail disruption to social resources and roles, switching the locations of life and work, plus the disruption caused by the fact that such moves are often provoked by dissatisfaction with current work or life states, or by the lure of more attractive work or life situations elsewhere. The research on geography and entrepreneurship is limited (Plummer, 2010). Whether the lure of distant business opportunities is stronger than the attraction of remaining in a familiar social context has not been studied explicitly. Dahl and Sorenson (2009) investigate the prevalence of geographic moves for business start-ups in Denmark in the period 1990-2007 and find that individuals seldom move in order to start a business, and that more often a change of location is related to being closer to family. A study by Pailhé and Solaz (2008) shows that in regulated labour markets, such as those in most of continental Europe, the benefits from a change of location are very limited, especially for the partner in a couple, who is often a “tied mover”. In our modelling of entrepreneurial life course events, I do not assume entrepreneurship needs to be a primary motive for geographic mobility. Being uprooted and relocated in a new social and economic context for reasons in part beyond ones‟ control, such as one‟s partner being offered a more attractive and better paid job, might provide the motivation for consideration of an alternative career path, such as entrepreneurship. I hypothesize, therefore, that: H4a: Moving to a different geographic area will positively affect an individual’s probability of entry into entrepreneurship. H4b: The greater the distance of past geographic mobility, the higher will be the individual’s probability of entering entrepreneurship. H4c: Moving with a partner to a different geographic area will positively affect an individual’s probability of entering entrepreneurship. 9

Methodology The dataset used to test our model is a combination of two archives maintained by Statistics Sweden, the official bureau of census. The first is LISA, which provides data on the household composition and demographics of all Swedish residents aged 16 and over, for 1989 to 2007. The second is the “cross-generational archive” which provides information on the labour market activities of individuals in the same household, for 1960 to 2002. These data provide several advantages. First, the long time period of the data directly addresses the challenge in studying life course events which evolve over time (Blossfeld, 1986; Giele & Elder, 1998). Second, the comprehensive nature of archival data enables us to address the problems of retrospective bias and outcome bias common in both life course and entrepreneurship research. Third, the setting of a small Nordic country characterized by relatively low rates of entrepreneurship – especially necessity entrepreneurship (Reynolds et al., 2003) – and a relatively gender egalitarian labour market situation is ideal for testing the model compared to settings characterized by a gendered labour market structure. A problem that emerges from the entrepreneurship literature is that new firms and their founders are very heterogeneous, ranging from “mom-and-pop” retail stores to venture-capital backed start-ups (Davidsson, 2004; Shane, 2008). To limit observed and unobserved sources of heterogeneity, I construct a risk set based on all individuals aged between 20 and 65 with at least a high school degree, who took up a new job in 1990, outside of the farming sector.2 The age limits reduce problems of geographic mobility and family choice, which are rare among individuals aged below 16; a focus on the newly employed eliminates problems with left-censoring, which occurs when a person is at risk of switching prior to our ability to observe them (Folta et al., 2010). In 1990, 565,820 individuals took up a new job and were at risk of transitioning

2

In LISA, self-employment farmers and fishermen are noted with occupational codes that makes them distinguishable from other types of self-employment.

10

from their current jobs to entrepreneurship. They remain in the sample until they die, or emigrate, or the observation period ends in 2007. Testing our theoretical model involves testing the constitutive and interactive nature of the theoretical variables derived from capital theory (Bourdieu, 1977) and life course theory (Blossfeld, 1986; Carroll & Mosakowski, 1989) for entrepreneurial processes. I choose to model two distinct events that have been investigated in depth in entrepreneurship research: the acts of creating and exiting a new venture. Hence, I seek to communicate with established streams of research by validating the theoretical importance of considering not only the more static factors associated with capital theory, but also the more serendipitous factors related to life course theory as they relate to entrepreneurial events. The two events of interest are those related to entrepreneurial entry and exit. Following earlier research, the first dependent variable, entry, is a dichotomous indicator of whether an individual in the relevant population was active in entrepreneurship during year t, but not in the preceding year. I define an entrepreneur as an individual who reports capital income from a company in which he or she work at least part time, and in which he or she has a majority ownership stake (cf. Folta et al., 2010). The second dependent variable, exit, is a yearly indicator of whether an individual leaves the entrepreneurial venture during any of the subsequent years (DeTienne, 2010). Independent variables: Geographic mobility. LISA includes information on county of residence, year variant. It also includes information on how often addresses are updated. I used this information to create two variables. The first, number of moves, is an ordinal scaled variable counting the number of times an individual moves during a focal calendar year. The second variable, mobility, measures the accumulated distance that an individual has moved during the whole period of observation, using a variant of the geographic dispersion index developed by

11

Sorenson and Audia (2000). A higher value of this measure implies high geographic mobility; lower values imply low mobility. Children. I measure the time-variant ordinal variables birth of children in household. Cultural Capital. I include an ordinal scaled variable measuring family firm background in entrepreneurship during upbringing. This variable is derived from LISA and Statistics Sweden‟s cross-generation database papper, which provide data on the labour market activities of all Swedish residents living in the same household from 1960 to 2002. The variable measures every year that parents where active in entrepreneurship during 1960, 1970, 1980, and 1986-1989. Married. I used information on household composition from LISA to construct a dummy variable that equals 1 if the individual becomes married or becomes cohabitant. General human capital – Years of education: I include a variable for years of education, the most common operationalization of general human capital in the literature (Arum & Muller 2004; Brüderl et al., 1992). Partner’s Specific human capital – Entrepreneurial experience: I used data from LISA to create the variable “entrepreneurial experience”, measured by the number of years of entrepreneurial experience during an prior career. By our research design, this variable is unobserved for all focal individuals since I wish to lower unobserved heterogeneity. Hence, experience is measured only exists for those marrying or becoming cohabitant. Ethnic background. Immigrants are known to have higher probabilities of entering entrepreneurship (Arum & Muller, 2004). I measure country of birth using dummy variables for groups that were found, in exploratory research to be related: Nordic, Western European/North American, Eastern European, other. 12

Partner’s income is measured in logarithmic form to control for the fact that household resources are often pooled in entrepreneurial ventures (Lin et al., 2000). I also include a control for household wealth to control for the possibility that financial slack/liquidity constraints might affect the probability of entry or exit (Holtz-Eakin et al. 1994). Age: All individuals living in Sweden receive a personal identification number based on their date of birth. I used this information to calculate individual‟s age (in years). Gender: Much of the existing research indicates that women entrepreneurs have lower entry rates and higher exit rates than male entrepreneurs (e.g. Arum & Muller, 2004). The personal identification number mentioned above captures the gender of the individual (men=1). To investigate the contingent nature of gender on our hypothesized predictor variables, I estimate separate models for men and women. Entrepreneurial earnings: It is difficult to disentangle the role of performance, entry and exit, since performance is difficult to ascertain for discontinued ventures. In the extant research, performance is either ignored or measured indirectly as “financial leverage” (Bates, 1990), or “money taken out of the business” (Gimeno et al., 1997). These are highly imperfect measures since some entrepreneurs choose to forego current benefits in order to reinvest in their business. The multi-level nature of our data allow us to overcome this problem since I can construct firm-level performance variables according to Hamilton‟s (2000) definition as [revenues – expenses = money taken out + retained earnings] in logarithmic form. Modelling approach Following the sociological research on careers, I use hazard models to test the constitutive and interactive nature of our theoretical variables on the two outcome variables of entry and exit (Sørensen, 2007). Since the data are annual, our models are estimated on the piecewise exponential distribution and eliminate the need for specific assumptions with 13

regard to shape of the hazard function. To allow the hazard to vary between years, the model is divided into yearly intervals with coefficients that are updated annually (Blossfeld & Rohwer, 1995). To test the contingent hypothesis relating life course events to capital resources, moderator or interaction effects are constructed between the ordinal variables of education and industry experience, and between children, marriage and geographic mobility. I follow the analytical procedures proposed by Brambor and colleagues (2006). In all models I therefore first present the main effects, with separate models estimated for men and women, followed by the interaction effects. Given the large sample size and risk of very weak and yet not very meaningful effects reaching statistical significance, all coefficients are displayed as hazard rates which eases interpretation of these as marginal effects, which I attend to for each hypothesis. All predictor and control variables were lagged one year to avoid simultaneity bias. Means, standard deviations and correlations are presented in Table 1. To guard against the potential of multicollinearity, variance inflation factors (VIF) were computed, but found to be below the generally accepted critical value of 10. All variables are presented in Table 1. -----------------------------------------Insert Tables 1, 2 & 3 here -----------------------------------------Results Table 2 shows piecewise exponential hazard models for entrepreneurial entry. Model 1 shows the direct effects; model 2 includes the interaction effects. Hence, I use model 2 to test all our hypotheses. I examine first the hypothesized effects of cultural capital in the form of parents‟ entrepreneurship, on entrepreneurial entry; the findings in Table 2 show that the coefficients for both men‟s and women‟s cultural capital in the base and interaction models are positive and significant at the 0.01 level or above. This confirms hypothesis 1a. Given the exhaustiveness of this dataset which tracks potential entrepreneurs over time, a question of at 14

least equal interest as significance testing is how meaningful are these effects in terms of size? The hazard coefficients of 1.022 for men‟s cultural capital and 1.009 for women‟s cultural capital in the respective interaction models indicate that each year that parents were active in entrepreneurship during the period overlapping the childhood years of the individuals in this study, increase the probability of entry by 2.2% for men and 0.9% for women. Of the 13,265 individuals that entered into entrepreneurship during the study period, 15.9% were from families that were entrepreneurially active during at least one year in the period 1960-1989. The mean value of cultural capital for this group of entrepreneurs who followed in their parents‟ footsteps is 4.6 for men and 3.6 for women, indicating that the mean marginal effect of cultural capital for men(women) that entered entrepreneurship and were from a family firm background, was 10.1%(3.2%). In the context of a modern economy, these findings indicate quite strong patterns of cultural capital. Looking briefly at the effect of cultural capital on entrepreneurial exit presented in Table 3, similar to Sørensen (2007) I find that cultural capital has a quite small effect on duration in entrepreneurship for men, significant at the 10% percent level, and not at all for women. Our next hypothesis posits that the positive effect of parents‟ entrepreneurship is contingent on the social structure of society and will be stronger for individuals from minority cultural backgrounds. I examine this hypothesis by pooling all minority dummies into a composite dummy variable which I interact with the cultural capital variable and enter in the interaction models. This variable, displayed in the bottom row of table 2, shows that the effect of parents‟ entrepreneurship for individuals from minority cultural backgrounds is strongly positive, in the expected positive direction, for both men (1.23, p>0.001) and women (1.09, p>0.01). The hazard coefficients of 1.23 and 1.09 for the respective dummy variables reveal that a minority cultural background increase the effect of parents‟ entrepreneurial background on entrepreneurship among men by 23% and women by 9%, which in our view are meaningfully large effect sizes.

15

If I consider the hypothesized effects on marriage and entrepreneurial entry, I can argue that marriage or cohabitation facilitates the integration of lives and careers, potentially men‟s and women‟s career prospects, and the differential ways that they adapt to/learn from each other. In the model, this is measured by interacting the dummy variable for marrying/cohabiting with the ordinal variable measuring partner‟ current or prior entrepreneurial experiences.3 The last row in Table 2 shows that the interaction variable measuring marrying/cohabiting a partner with current or prior entrepreneurial experience is positive and significant for both men (1.081, p>0.05) and women (1.09, p>0.05), which supports hypthesis 2a but not hypothesis 2b. The hazard coefficients of 1.081 and 1.090 for the respective variables reveal that marrying/cohabiting with a man with current or prior entrepreneurial experience positively enhances women‟s probability of entry into entrepreneurship by a factor of 0.8% for each year of experience of the male partner, and enhances men‟s probability of entry by a factor of 0.9% for each year of experience of the wife/partner.4 These are meaningful effects albeit not as strong as those related to human and cultural capital. The rejection of hypothesis 2b indicates that, contrary to what hypothesized, both women and men are affected by the entrepreneurial experience of their partners. I next investigate perhaps one of the strongest accentuated variables in life course theory which takes a central role in our theory of the entrepreneurial life course: the hypothesized effects of children on entrepreneurial entry. Looking at the coefficients for number of children in Table 2, I can see that having children positively influences the likelihood of a man entering entrepreneurship (1.058, p>0.001), and also positively influences

3

Since the availability of a partner‟s specific human capital is obviously conditional on a focal individual having a partner, the constitutive term for specific human capital is not part of the base model. 4

Note that while same-sex couples are potentially included in these models, the phrasing of hypotheses or interpretations is not contingent on the sex of the partners in a household. From the perspectives of both life course theory and gender theory, it is possible that the effects of the partner‟s human capital is distinct between same-sex and different-sex households. This is a subject for future research.

16

women‟s likelihood of entry (1.038, p>0.05). This supports hypotheses 3a and 3b. The hazard coefficients of 1.057 and 1.038 for women indicate that each additional child raises the likelihood of entry by 5.7% for men and 3.8% for women. After focusing mainly on variables related to various sources of human and cultural capital resources and their interaction with life course events, I next examine the role of financial capital on life course by investigating the hypothesized effects of partners‟ income from paid employment, on entrepreneurial entry. Table 2 presents the coefficients for partners‟ income which are positive and in the expected direction (1.070, p>0.001 and 1.097, p>0.001, for men and women, respectively). Interpretation of the marginal hazard coefficients is less meaningful in that income is measured in log format, but the effect is fairly strong. The final set of analyses on how life course events shape entrepreneurial entry, pertains to the effects of geographic mobility on entrepreneurial entry. The first two rows in Table 2 show that Number of moves is positive and significant in the expected direction (1.040, p>0.001 and 1.032 p>0.01, for men and women respectively) while mobility (3rd and 4th rows) is not significant in any of the model specifications. This indicates that the actual act of moving from one location to another not the actual distance moved, is the source of geographic life course events that matters for entrepreneurial entry. The marginal hazard rates indicate that each geographic move during the year preceding entry increases the probability of entry by 4.0% for men and 3.2% for women, which is a quite substantial effect. Our final hypothesis for geographic sources of entrepreneurial entry stated that moving with a partner to a different geographic area should positively enhance an individual‟s probability of entry. This is tested by a variable for „number of moves during past year‟ interacted with married/cohabitant status during the previous year. The interaction variable, last row in Table 2, is weakly significant for men (1.017, p>0.10) but stronger for women (1.027, p>0.05). This indicates that in addition to the effect of relocating on the probability of entering 17

entrepreneurship, moving with a partner increases the probability further by a factor of 1.7% for men (subject to debate on the weak level of statistical significance) and 2.7% for women, for each move during the previous year. The second part of the empirical examination thus concerns the test of hypotheses related to exit. Table 3 shows piecewise exponential hazard models for entrepreneurial exit. Again, model 1 shows the direct effects while model 2 includes the interaction effects. I examine the hypothesized effects of children and entrepreneurial exit. The coefficient for number of children (Table 3) reveals that having children lowers the probability that a man will exit from entrepreneurship (0.948, p>0.01). However, columns 4 and 5 in Table 3 show that having children does not influence women‟s probability of exit in either direction (0.968, p>0.26), which refutes our hypotheses and earlier research suggesting that the presence of children in the household will increase the probability of a women exiting entrepreneurship.5 The marginal hazard rates indicate that each additional child decreases the probability of exit by 5.2% for men, which is a quite substantial effect. Conclusions In this paper I posited a model of entrepreneurship based on life course theory and capital theory. I used these theories to derive a number of testable prediction, some of which I tested against a longitudinal dataset comprising individual life courses in a random sample of 565,820 individuals. Despite the tentative evidence providing support for the life course model of entrepreneurship, this is a model in formation and would benefit from further theoretical development. A potential objection to the life course perspective is that it constitutes merely a sociological application of an even more higher-order framework or loosely coupled social systems where enterprising individuals and entrepreneurial opportunities are in a constant state of flux, such as in March and Olsen‟s (1978) “garbage 5

These effects are conditional on controlling for entrepreneurial earnings, as per the phrasing in our hypotheses.

18

can model” or Bygrave‟s (1989) model of “entrepreneurial trigger events”. While there are some similarities, I believe that our micro-level model of the entrepreneurial life course is distinct from organizational oriented models such as the garbage can or Bygrave‟s work in at least two respects: First, in contrast to these models, an integrated part of our model is that triggering events may be exogenous to the entrepreneurial process. Life course theory suggests that personally critical events, such as childbirth, divorce, or migration, might both (i) inspire an individual to considering engaging in entrepreneurship; and (ii) lead to an upheaval in the social fabric that triggers an individual already considering entrepreneurship to “take the plunge”. Second, the current model provides a more micro-oriented framework that complements the more higher-order approach provided by the garbage can model and Bygrave‟s model of trigger events. Another line of potential criticism is that management oriented research may overlook the benefits of a structured view of the entrepreneurial processes such as the one outlined on this model of entrepreneurial life course, in that it might seem less applicable for deriving managerial strategies. It is important to note here that our framework does not depend solely on fixed idiosyncrasies, nor does it dispense with individual agency. Idiosyncratic knowledge, endogenously enacted opportunities, and the willingness of the individual to take the plunge and become an entrepreneur once this potential is conceived, are integral parts of our model: individual‟s entrepreneurial choices can be the result of chance, good timing, or simply being in the right place at the right time. Further extensions to our theoretical model could strengthen as well as challenge its potential for modelling the entrepreneurial processes. For example, recent advances in social network theory suggest integrating the notion of “network content” and “network decay” (Burt, 2000) in models relying on social capital, such as the one outlined here, to allow for a more socialized view of entrepreneurial agency. This would

19

however require detailed agent-centric network data collected longitudinally that is matched to other characteristics of individuals‟, requiring vast resources for data collection. In short, this paper outlines and empirically test a model of entrepreneurial choices based on life course theory and capital theory. While I found support for most of the hypotheses derived from the theoretical frameworks, some limitations and gaps remains to be developed by further research. I would welcome further contributions along these lines. I believe that the general framework outlined in this paper contributes to a dynamic view of the entrepreneurial processes as being shaped at the intersection between resourceful individuals and serendipitous opportunities, which are neither planned nor optimized, but when wilfully pursued lead individuals to create new firms, potentially spreading new ideas and innovations, and in the long run contributing to the creation of new industries.

20

References Aldrich, H. 1979. Organizations and Environments. Englewoods Cliffs, NJ: Prentice-Hall. Aldrich, H. E., & Kenworthy, A. L. (1994). The Accidental Entrepreneur: Campbellian Antinomies and Organizational Foundings. In J. A. C. Baum & J. V. Singh (Eds.), Variations in Organization Science: In Honor of Donald t. Campbell. . Arum, R., & Muller, W. 2004. The Re-Emergence of Self-Employment: A Comparative Study of Self-Employment Dynamics and Social Inequality. Princeton, NJ: Princeton University Press. Baker, T., Miner, A. S., & Eesley, D. T. 2003. Improvising firms: Bricolage, account giving and improvisational competencies in the founding process. Research Policy, 32(2): 255. Bates, T. 1990. Entrepreneur human capital and small business longevity. The Review of Economics and Statistics, 72(4): 551-559. Boden, R. J., & Nucci, A. R. 2000. On the survival prospects of men's and women's new business ventures. Journal of Business Venturing, 15(4): 347-362. Bourdieu, P. 1977. Cultural reproduction and social reproduction. In J. Karabel, & A. Halsey (Eds.), Cultural reproduction and social reproduction. New York: Oxford University Press. Brush, C. G. 1992. Research on women business owners: Past trends, a new perspective and future directions. Entrepreneurship Theory & Practice, 16(4): 5-30. Bygrave, W., 1989. The Entrepreneurship Paradigm (2). Entrepreneurship Theory and Practice 14, pp. 7–26 Blossfeld, H.-P. 1986. Career opportunities in the Federal Republic of Germany: a dynamic approach to the study of life course, cohort, and period effects. European Sociological Review, 2(3): 208-225. Blossfeld, H.-P., & Rohwer, G. 1995. Techniques of Event History Analysis: New Approaches to Causal Analysis. Mahwah, New Jersey: Lawrence Erlbaum Associates. Boden, R. J. 1999. Gender inequality in wage earnings and female self-employment selection. Journal of Socio-Economics, 28(3): 351-364. Brambor, T., T.W. Clark, M. Golder. 2006. Understanding interaction models: Improving empirical analysis. Pol. Analysis 14 63-82. Bruce, D. 1999. Do husbands matter? Married women entering self-employment. Small Business Economics, 13(4): 317.

21

Brüderl, J., Preisendörfer, P., & Ziegler, R. 1992. Survival Chances of newly founded business organizations. American Sociological Review, 57(2): 227-242. Brush, C. G. 1992. Research on women business owners: Past trends, a new perspective and future directions. Entrepreneurship Theory & Practice, 16(4): 5-30. Burt, R. S. 2000. The network entrepreneur. In R. Swedberg (Ed.), Entrepreneurship: The Social Science View: 281-307. Oxford:England: Oxford University Press. Butler, J. S., & Herring, C. 1991. Ethnicity and Entrepreneurship in America - toward an Explanation of Racial and Ethnic-Group Variations in Self-Employment. Sociological Perspectives, 34(1): 79-94. Carr, D. 1996. Two paths to self-employment? Women's and men's self-employment in the United States, 1980. Work and Occupations, 23(1): 26-53. Carroll, G. R., & Mosakowski, E. (1987). The Career Dynamics of Self-Employment. Administrative Science Quarterly, 32(4): 570-589. Cooper, A. C. 1985. The role of incubator organizations in the founding of growth oriented firms. Journal of Business Venturing, 1: 75-86. Connelly, R. 1992. Self-Employment and Providing Child-Care. Demography, 29(1): 17-29. Cornelius, B., Landstrom, H., & Persson, O. 2006. Entrepreneurial studies: The dynamic research front of a developing social science. Entrepreneurship Theory and Practice, 30(3): 375. Dahl, M. S., & Sorenson, O. 2009. The Embedded Entrepreneur. Entrepreneurs‟ location choices and the performance of their new ventures. European Management Review. 6(3): 172-181. Davidsson, P. 2004. Researching Entrepreneurship. New York: Springer. Davidsson, P., & Honig, B. 2003. The role of social and human capital among nascent entrepreneurs. Journal of Business Venturing, 18(3): 301-331. DeTienne, D. 2010. Entrepreneurial exit as a critical component of the entrepreneurial process: Theoretical development. Journal of Business Venturing, 25(2): 203-215. Douglas, E., J., & Shepherd, D. A. (2000). Entrepreneurship as a utility maximizing response. Journal of Business Venturing, 15: 231-251. Drobnic, S., Blossfeld, H.-P., & Rohwer, G. 1999. Dynamics of Women's Employment Patterns over the Family Life Course: A Comparison of the United States and Germany Journal of Marriage and the Family, 61(1): 133-146.? Not in text Elder,G.H (Ed) (1985). Life course dynamics: Trajectories and transitions, 1968-1980. Ithaca: Cornell University Press. 22

Elder, G.H (1994). Time, Human Agency, and Social Change: Perspectives on the Life Course. Social Psychology Quarterly, 57(1): 4-15. Elder, G.H (1998). The Life Course as a Developmental Theory. Child Development, 69(1) 112. Evans, D.S. & Jovanovic, B. (1989). An Estimated Model of Entrepreneurial Choice Under Liquidity Constraints. Journal of Political Economy, 97(4): 808-827. Folta, T. B., Delmar, F., & Wennberg, K. 2010. Hybrid Entrepreneurship. Management Science, 56(2): 253-269. Giddens, A. (1984). The constitution of society: Outline of the theory of structuration. Berkeley:University of California Press Giele, J., & Elder, G. (Eds.). 1998. Methods of life course research: Qualitative and quantitative approaches. Thousand Oaks, CA: Sage. Gimeno, J., Folta, T. B., Cooper, A. C., & Woo, C. Y. 1997. Survival of the fittest? Entrepreneurial human capital and the persistence of underperforming firms. Administrative Science Quarterly, 42(4): 750-783. Glass, J., & Fujimoto, T. 1995. Employer Characteristics and the Provision of Family Responsive Policies. Work and Occupations, 22(4): 380-411. Hamilton, B. H. 2000. Does entrepreneurship pay? An empirical analysis of the returns of self-employment. Journal of Political Economy, 108(3): 604-631. Hao Z, Seibert SE, Hills GE (2005) The Mediating Role of Self-Efficacy in the Development of Entrepreneurial Intentions. Journal of Applied Psychology, 1265-1272. Heckhausen, J. and R. Schulz 1995. A life-span theory of control. Psychological Review, 102(2): 284-304. Holtz-Eakin, D., Joulfaian, D., & Rosen, H. S. 1994. Sticking it out: Entrepreneurial survival and liquidity constraints. The Journal of Political Economy, 102(1): 53. Kalleberg, A. L., & Leicht, K. T. (1991). Gender and organizational performance: Determinants of small business survival and success. Academy of Management Journal, 34(1): 136-161. Kim, P. H., Aldrich, H. E., & Keister, L. A. 2006. Access (not) denied: The impact of financial, human, and cultural capital on entrepreneurial entryin the United States. Small Business Economics, 27(1): 5-22. Krueger, N. F., Reilly, M. D., & Carsrud, A. L. 2000. Competing models of entrepreneurial intentions. Journal of Business Venturing, 15(5-6): 411-432.

23

Lin, Z., Picot, G., & Compton, J. 2000. The Entry and Exit Dynamics of Self-Employment in Canada. Small Business Economics, 15: 195-125. Macpherson, D. A. 1988. Self-Employment and Married-Women. Economics Letters, 28(3): 281-284. March, J., & Olsen, J. 1978. A garbage can model of organizational choice. Administrative Science Quarterly. Meyer, M.H. (1996). Making Claims as Workers or Wives: The Distribution of Social Security Benefits. American Sociological Review, 61(3): 449-465. Padgett, J., & Ansell, C. 1992. Robust action and the rise of the Medici. American Journal of Sociology, 98: 1259-1330. Pailhé, A. & Solaz, A. (2008). Professional Outcomes of Internal Migration by Couples: Evidence from France. Population, Space and Place, 14:347-363 Papalia, D.E. & Olds, S.W. (1981). Human Development. New York: McGraw-Hill. Plummer, L. A. 2010. Spatial Dependence in Entrepreneurship Research: Challenges and Methods. Organizational Research Methods, 13(1): 146-175. Reynolds, P. D., Bygrave, W., Autio, E., Cox, L. W., & Hay, M. 2003. Global Entrepreneurship Monitor 2003 Executive Report. Babson: MA: Babson College, Ewing Marion Kauffman Foundation, London Business School. Shane, S. (2008). Illusions of Entrepreneurship. New Haven: Yale University Press. Shapero, A. 1982. Social Dimensions of Entrepreneurship. In C. A. Kent, D. L. Sexton & V. K. H. (Eds.), Encyclopedia of Entrepreneurship (pp.72-90). Englewood Cliffs, NJ Prentice-Hall. Schroots, J. J. F. 2003. Life course Dynamics. European Psychologist, 8(3): 192-199. Sørensen, J.B. (2007). Closure and Exposure: Assessing Alternative Mechanisms in the Intergenerational Inheritance of Self-employment. Research in the Sociology of Organizations: The Sociology of Entrepreneurship, pp. 12-35: JAI Press. Sorenson, O., & Audia, P. G. 2000. The social structure of entrepreneurial activity: Geographic concentration of footwear production in the United States, 1940-1989. American Journal of Sociology, 106(2): 424-462. Steinmetz, G., & Wright, E. O. 1989. The Fall and Rise of the Petty Bourgeoisie - Changing Patterns of Self-Employment in the Postwar United-States. American Journal of Sociology, 94(5): 973-1018. Uggen, C. 2000. Work as a Turning Point in the Life Course of Criminals: A Duration Model of Age, Employment, and Recidivism American Sociological Review, 65(4): 529-546. 24

Table 1 Variable mean values and correlation matrix Variable

Mean St. dv.

1. Entry

0.023 0.151

2. Exit

0.004 0.354 0.000

3. Number of moves

0.367 0.705 0.023 0.002

4. Mobility

6.60

5. Children

0.974 1.108 0.034 -0.026 0.000 0.044

6. Cultural Capital

1)

2)

3)

4)

5)

6)

7)

8)

9)

10)

12)

13)

14)

18)

1.016 0.012 0.046 -0.002

0.269 1.007 0.019 -0.043 7. Gen. Human capital 12.33 2.313 0.076 -0.098 8. Married/Cohabitant 0.723 0.323 0.016 -0.074 9. Ethnicity: Nordic 0.074 0.262 -0.004 0.014

-0.002 0.010 -0.021 0.021 0.024 0.005 0.090 0.113 0.011 0.029 0.089 0.002 -0.037 -0.011 -0.017 0.004

10. Ethnicity: Western

0.026 0.158 0.015 -0.002 -0.002 0.039 -0.003 -0.006 -0.013 -0.024

11. Ethnicity: Eastern

0.018 0.132 0.016 0.002 0.001 0.024 0.005 -0.001 -0.010 0.002 -0.029

12. Ethnicity: Other

0.032 0.178 -0.009 0.002 0.003 0.070 0.018 -0.013 -0.019 -0.002 -0.030 -0.017

13. Spouse salary(log) 11.04 1.180 0.042 -0.107 0.000 -0.086 14. Wealth(log) 60.93 4.233 0.094 -0.041 -0.001 0.000 15. Age 33.61 11.102 0.063 0.020 0.000 -0.013 16. Gender 0.509 0.499 0.023 -0.022 0.000 0.001 17. Ent. earnings(log)

11)

-0.048 0.044 0.020 0.032 -0.041 -0.024 -0.025 0.024 0.079 0.041 0.838 -0.006 -0.008 -0.024 -0.091 0.020 -0.083 -0.032 0.004 -0.004 0.005 0.002 -0.015 0.015 0.000 0.005 0.037 0.124 0.061 0.044 0.042 -0.042 0.212 0.085

10.12 1.572 0.010 -0.313 -0.002 -0.017 0.007 -0.008 -0.004 -0.001 -0.025 0.018 -0.009 0.041 0.210 0.153 0.039

25

Table 3: Piecewise exponential models on entrepreneurial entry Men Interaction Variable Base model model Number of moves 1.052*** 1.040** (0.012) H4a: (0.010) Mobility 1.099 1.125 (0.080) (0.082) Children 1.060*** 1.058*** (0.009) H3a/ H3b: (0.008) Cultural Capital 1.038** 1.022** (0.002) (0.004) H1a: Human capital 1.072** 1.042* (0.006) (0.018) Married/Cohabitant 1.045 1.035 (0.043) (0.043) Ethnicity: Nordic 1.008*** 1.007*** (0.002) (0.002) Ethnicity: Western 1.020*** 1.016*** (0.003) (0.003) Ethnicity: Eastern 0.991 0.993 (0.016) (0.016) Ethnicity: Other 0.858*** 0.886** (0.036) (0.038) Partner’s income(log) 1.064*** 1.070*** (0.022) (0.022) 3e: Wealth(log) 1.024*** 1.022*** (0.005) (0.005) Age 1.007+ 1.002 (0.004) (0.004) Any Ethnicity X Cultural Capital 1.230*** (0.004) H1b:+ Marriage X Specific Human Capital 1.081* H2a:0 (0.040)

Women Interaction Base model model 1.033** 1.032** (0.008) (0.008) 1.099 1.125 (0.080) (0.082) 1.040* 1.038* (0.010) (0.012) 1.013** 1.009** (0.001) (0.001) 1.060* 1.030 (0.017) (0.020) 1.050 1.035 (0.038) (0.033) 1.010*** 1.009*** (0.001) (0.001) 1.030*** 1.020** (0.004) (0.005) 0.998 0.994 (0.012) (0.014) 0.865*** 0.890** (0.038) (0.040) 1.104*** 1.097*** (0.010) (0.012) 1.008*** 1.007*** (0.002) (0.002) 1.006 1.005 (0.004) (0.004) 1.009*** (0.002) 1.090* H2b:+ (0.008)

Number of moves X married Time dummies Log. Likelihood ∆ Log-likelihood: # Entries: # Unique Cases:

1.017+ 1.027* (0.010) H4c:+ (0.009) Yes Yes Yes Yes -6724.138 -6619.288 -5882.928 -5790.124 182.58*** 56.53*** 7,674 5,591 288,250 277,570

Note: Coefficients in hazard rate format, no constant estimated. All models include time dummies. ΔLR is for improvement in model fit versus base model. + p <.10, * p <.05, ** p <.01, *** p <.001. (two-tailed).

Table 3: Piecewise exponential models on entrepreneurial exit Men Women Interaction Interaction model Base model model 0.997 0.996 0.990 (0.012) (0.007) (0.007) 1.125 0.920 0.922 (0.082) (0.069) (0.068) 0.948** 0.965 0.968 (0.024) H3d: (0.162) (0.174) 0.992+ 0.976 0.995 (0.003) (0.091) (0.093) 0.967*** 0.957*** 0.950** (0.056) (0.002) (0.004) 1.035 1.001 1.002 (0.043) (0.001) (0.001) 1.007*** 0.990** 0.993** (0.002) (0.000) (0.000) 1.016*** 0.982** 0.984** (0.003) (0.003) (0.004) 0.993 0.950 0.951 (0.016) (0.031) (0.031) 0.886** 0.981*** 0.980*** (0.038) (0.001) (0.001) 0.970*** 0.989** 0.934** (0.022) (0.037) (0.030) 1.010 1.005 1.006 (0.025) (0.014) (0.013) 1.002 1.030 1.033 (0.004) (0.089) (0.090) 0.920*** 0.952*** 0.952*** (0.019) (0.016) (0.019) 0.968 0.808 (0.090) (0.120) Yes Yes Yes Yes -2860.572 -2819.2828 -2392.039 -2385.120 13.18** 23.12** 5,265 3,853 7,674 5,591

Base Variable model Number of moves 0.994 (0.012) Mobility 1.099 (0.080) Children 0.936** H3c: (0.02) Cultural Capital 0.992+ (0.003) Human capital 0.972*** (0.006) Married/Cohabitant 1.045 (0.043) Ethnicity: Nordic 1.008*** (0.002) Ethnicity: Western 1.020*** (0.003) Ethnicity: Eastern 0.991 (0.016) Ethnicity: Other 0.858*** (0.036) Spouse salary(log) 0.964*** (0.022) Wealth(log) 1.024 (0.025) Age 1.007+ (0.004) Entrepr. earnings(log) 0.920*** (0.019) Number of moves X married Time dummies Log. Likelihood ∆ Log-likelihood: # Exits: # Unique Cases:

Note: Coefficients in hazard rate format, no constant estimated. All models include time dummies. ΔLR is for improvement in model fit versus base model. + p <.10, * p <.05, ** p <.01, *** p <.001. (two-tailed).

27