Regional Influences on the Emergence of Family and Non-Family Businesses Miriam Bird Karl Wennberg Center for Entrepreneurship and Business Creation Stockholm School of Economics Box 6501 SE-113 83 Stockholm, SWEDEN. Phone: +46 8 736 9354 Email: [email protected] Abstract: This paper integrates insights from economic geography and organizational ecology to the entrepreneurship field by constructing a theoretical framework that theorizes how the regional bed for start-ups may affect family firms and non-family firms in differential ways. Using a rich multi-level data set, we investigate how characteristics of the economic milieu of regions influence firm births. We find that economic factors such as regional wealth and number of children born strongly affect the number of start-ups of non-family firms. However, the number family start-ups is more strongly tied to the level of small businesses and the political regulatory regime within the municipality. Taken together, our findings indicate that family start-ups are more susceptible to the local non-economic context than non-family start-ups.

Introduction This paper is posited at the intersection of economic geography, organizational ecology and entrepreneurship research, in investigating regional factors that determine the emergence of family versus non-family businesses. Research has argued that entrepreneurs are inseparably linked to their family (Dyer, and Handler 1994; Aldrich, and Cliff 2003) but research still lacks knowledge if and how the emergence, growth, and survival of family businesses differ systematically from that of non-family businesses (Brockhaus 1994; Zahra et al. 2004). Yet, we know that a substantial part of all start-up companies are founded by families but exactly how prevalent this is in modern economies, is still rather unknown in the literature. Families are essential for entrepreneurial endeavours since family members share a common identity, have strong mutual bands of trust, and often have opportunities to discuss business ideas. Although research indicates that family businesses are driven by different dynamics than non-family businesses (Nordqvist, and Melin 2010), few studies have contrasted these two types of

businesses (Brockhaus 1994; Zahra et al. 2004). The studies in existence have primarily attended to differences between family and non-family businesses on the individual or firm-level, emphasizing the distinctive capabilities, resources, and features that family businesses possess as opposed to non-family businesses (Zahra et al. 2004; Littunen, and Hyrsky 2000; Naldi et al. 2007). However, no study to date has examined how environmental characteristics may nurture the emergence of family and non-family businesses differently. Such comparative studies may help in understanding how the family shapes the economic characteristics of a company and how this in turn affects outcomes on the regional level. This is a deplorable gap in the literature since we know that family businesses are strongly culturally and regionally embedded (Colli et al. 2003) and furthermore, that the level of enterprising activities in a region is an important determinant of regional economic vitality (for instance Coffey, and Polèse 1985; Malecki 1993). In this paper we posit that certain economic and social factors may affect the emergence of family and non-family businesses differentially. Hence, our study addresses an interesting theoretical and empirical gap in the literature since both ecological theory and family business research posit that start-up processes are strongly characterized by the people and social environment that surround them (Romanelli 1991). Thus, the social embeddedness of economic activities becomes especially important. We suggest that a range of social and environmental factors influence the formation of family and non-family businesses differently. Specifically, we theorize that regions characterized by lower income levels, a high proportion of ethnic minorities, situated in rural areas and with more conservative political inclinations will be characterized by a higher prevalence of family start-ups as opposed to non-family start-ups (Johnson, and Parker 1996; Bull, and Winter 1991; Gianetti, and Simonov 2007; Lee et al. 2004). Conversely, we argue that regions characterized by higher levels

economic demand and proximity to urban areas will have an increased number of non-family start-ups (Lee et al. 2004). To investigate these hypotheses we draw upon an unusually rich multi-level data material on all start-ups across the 290 Swedish municipalities between 1991 and 2006. Our empirical analysis focuses on how characteristics of the economic milieu of regions influence firm births. Our theoretical framework and the findings provide theoretical, empirical, and methodological contributions to research in entrepreneurship and economic geography. Specifically, we show that the patterns affecting the emergence of family businesses as opposed to non-family businesses differ systematically, substantiating the view that these different types of firms should be distinctly scrutinized in macro-oriented studies. On the micro level, our study supports the notion that the family system has a substantial influence on the business founded (Aldrich, and Cliff 2003; Nordqvist, and Melin 2010). Finally, since the importance of small and young enterprises is large in terms of new job generation (Birley 1986), this study helps to generate policy implications to address the needs of non-family businesses as well as of family businesses. Theory and Hypotheses A substantial amount of literature in entrepreneurship, population ecology, and economic geography suggests that geographic factors are important in shaping the evolutionary paths by which new entrepreneurial firms emerge. A common platform for these three strands of literature is that they are based on an evolutionary view of economic processes, in which the environment is a set of influences that allows some firm to emerge, survive or change forms (Aldrich 1990; Singh, and Lumsden 1990). This view is also supported by Aldrich and Ruef (2006) who postulate that organizations are strongly influenced by their context and environment. Hence, it could be argued that the environment embraces the social processes and conditions in which

potential founders find themselves in. In the following it will be discussed what the main essence of the three kinds of literature is and how they relate to our research topic. In organizational ecology the focus is primarily on the evolutionary perspective which posits that certain changes in populations occur due to selective demise and survival of organizations. The ecological approach emphasises the importance of resources available in a certain environment for the existence and emergence of new organizations. Since organizations are dependent on certain combinations of resources at hand, they respond in some way to the environmental forces and constraints in place. Firms compete for specific resources justifying their existence implying also an increasing specialisation and consequently leading to their positioning in special niches. This causes a state of competitive but also cooperative interdependences between organizations (Aldrich, and Ruef 2006). Available resources determine how many companies the environment can absorb and therefore also the number of possible startups. This process of adapting to a particular environmental setting and to resource constraints affects the survival and demise of organizations. Components often associated with organizational ecology are spatial location, community embeddedness (Ruef 2000, and proximity to urban areas (Pennings, 1982). Ecological theory suggests that regional start-up rates should increase with spatial density of similar organizations (Hannan et al. 1995; Aldrich 1990). The density of similar organizations determines to some extent how developed a population is. From this perspective, the emergence of businesses depends on the size, age and the legitimization of existing populations. An increasing density of similar organizations implies not only a higher availability of information but also the emergence of social networks (Aldrich, and Ruef 2006). These networks might give potential entrepreneurs the possibility to learn about potential opportunities. Although a developed population might offer favourable conditions to new

companies, one has to consider that once the population has reached a certain size, competition and resource constraints might hinder the entrance of new companies. The importance of geography as shaping business activities has been one of the strong emerging strands in economic research – especially in the “new economic geography” research. Research in economic geography has noted that the level of start-ups in regions often feeds into a self-reinforcing process that forms an agglomeration of related firms, cooperating and competing with each other (Feldman, Francis, and Bercovitz 2005). Components often associated with agglomeration economics are industry structures (i.e. industry size and differentiation), the density of people, and the quality of infrastructure. From this perspective, the emergence of businesses depends on the availability of an output market that may be served by new firms, and is facilitated by the availability of regional supply of resources such as financial and human capital required to set up a firm (Mason, and Harrison 2006; Pe‟er, and Vertinsky 2008). From a macro-oriented community perspective, the creation of a new firm is hence an important entrepreneurial act carried out by individuals, teams or families (Aldrich, and Ruef 2006). Following our perspective on the regional differences shaping patterns of family start-ups and non-family start-ups, we follow Gartner (1988) and Low and MacMillan (1988) in viewing entrepreneurship as the process by which new firms come into existence. Viewing entrepreneurship as the process by which new firms come into existence, the predictors or outcomes of such processes can be studied on different levels (Davidsson, and Wiklund 2001). Population-level studies of organizational emergence tend to investigate distinct “populations” of firms that rely on the same resource base and serve similar cause (Hannan et al. 1995; Singh, and Lumsden 1990). Hence, organizations belonging to one population share similar characteristics, features and have a common understanding of the socio-cultural rules in place. The next level, the community level, encompasses several populations. This level offers

populations the same environmental conditions and comprises both economic and relational aspects (Aldrich, and Ruef 2006). We argue that the municipality level comes close to the community level used in population ecology. This is in turn reflected in the competencies and routines prevalent in a certain municipality. In the below we draw upon the perspectives of economic geography, organizational ecology, and entrepreneurship to theorize around the roles of regional embeddedness and regional economies in shaping the conditions for family start-ups and non-family start-ups. Start-ups and Economic Demand Size The varying local economic demand conditions across municipalities are likely to have an impact on the emergence of business start-ups (Johnson, and Parker 1996). Economic demand is influenced by the local market conditions prevalent in a municipality. Two kinds of influences are likely to be important: the market size (in terms of population size) and the level of income. The importance of these two factors has been vindicated by several studies in examining firm creation in so-called “local market areas” (Johnson, and Parker 1996; Reynolds 1994). Level of income may affect firm emergence since higher level of income per capita reflects a higher level of regional, economic well-being (Reynolds 1994). It is undeniable that a profit-seeking entrepreneur will have other motivations to open a business than a family. We argue that for family businesses this aspect is less important than for non-family businesses since family businesses are more often driven by non-pecuniary motives (Gimeno 1997): Hypothesis 1a: Population size in a focal municipality will be of higher importance for nonfamily start-ups than for family start-ups. Hypothesis 1b: Population growth in a focal municipality will be of higher importance for nonfamily start-ups than for family start-ups. Hypothesis 1c: The level of income in a focal municipality is of higher importance for non-family than for family start-ups.

Start-ups and Urbanization Urbanization has for long been considered an important factor in explaining the birth of enterprises. Pennings (1982) who studies how urban conditions affect start-up activity in certain industries, posits that urban conditions foster entrepreneurial activity. Urban areas are believed to have a strong impact on the firm formation process due to the fact that cities provide access to strategic inputs through dense concentrations of both consumers and other companies (Johnson, and Parker 1996). As areas become more urban, the diversity of the organizations increases, this potentially attracts entrepreneurs to find and fill special niches (Aldrich, and Ruef 2006). Further, due to the agglomeration of knowledge and the likelihood of knowledge spillovers, the identification of business opportunities is higher in urban areas. Although urban areas might offer favourable conditions, entrepreneurs often start their company at the place of residence even though economic factors might be more favourable elsewhere (Dahl, and Sorenson 2009). It is common that potential entrepreneurs prefer to stay in the area familiar and known to them and therefore forego to optimize their cost benefit structure (Pennings 1982). This is particular noteworthy to compare with economic arguments for local choice, were firm births are believed to depend on the availability of resources in the form of economic inputs in combination with the existence of a large output market (Pe'er et al. 2006). Hence, while family firms are characterized by a higher degree of “socially embeddedness” than non-family businesses (Zahra et al. 2004), we believe they are also likely to be more “regionally embeddeded” than other firms. This indicates that family businesses should be more likely to prosper in rural region than non-family businesses, leading us to hypothesize: Hypothesis 2: Family start-ups are more likely to emerge in rural municipalities than non-family start-ups.

Start-ups and Immigration The rate of firm formation and the tendency to engage in family enterprises are known to differ across social and ethnic strata in society. In modern developed economies, the predominant differences concern different ethnicities. Immigrants often have limited opportunities and are excluded from the labour market due to lacking language skills, general high unemployment, and discrimination towards them (Stiles, and Galbraith 2003). Therefore, they seek other sources of income and often consider self-employment as an attractive possibility. Due to the strong personal ties and collective interests, the family facilitates the access to labour resources since family members such as spouses and children get involved in the family business. Second, the social, human and financial capital available within the family is often reverted to and thus facilitates the start-up process. However, it is noteworthy that these effects might vary between ethnic groups. Sanders and Nee (1996) have shown that family composition (i.e. having a spouse or teenage relatives) increases the probability of being self-employed among immigrants. Furthermore, the rate of self-employment among immigrants has increased in Sweden in recent years indicating the high importance of self-employment for immigrants (Hammarstedt 2004). Given the fact that the family is so strongly involved in the start-up process, we expect: Hypothesis 3: Family start-ups are more likely to appear in regions characterized by high number of immigrants than non-family start-ups. Start-ups and pre-existing Role Models Potential entrepreneurs who seek to establish a business where there are fewer small firms will have to accumulate new knowledge about the market and need to gain legitimacy for themselves (Hannan et al. 1995; Aldrich, and Ruef 2006). The validity of these patterns have been documented in research showing that regions with a high number of small firms generally have a high number of start-ups as opposed to regions being dominated by large companies (Garofoli 1994). The question of whether other small firms in the region provide access to knowledge and

legitimacy is of particular importance for family firms, since these are often highly regionally embedded (Anderson et al. 2005). With this we mean that the social structure of relations that tie economic actors together is highly spatial in nature. Individuals are influenced directly by those they interact with in the regional space (Johannisson, Ramirez-Pasillas, and Karlsson 2002), and also indirectly by those that they share common contacts with (Dobrev 2005). While direct ties provides access to both entrepreneurial role models and knowledge about how to start and run a new business, indirect ties also provide access to a wider array of role models in the region. Combined, these direct and indirect ties facilitate the creation and amalgamation of localized business networks that strengthen the „embeddedness‟ of local businesses (Feldman et al. 2005). Since family businesses are dependent on networks and existing knowledge compared to other entrepreneurs (Donckels, and Fröhlich 1991), we believe that the local embeddedness following from the density of small business in the region will be of particular importance to these firms (Singh, and Lumsden 1990). This leads us to believe that the effects of the density of small business in the region on new start-ups will be of particular importance for family businesses: Hypothesis 4a: The number of existing businesses will have a greater impact on the emergence of family start-ups than on non-family start-ups In modern economies, the service industry can be regarded as acting as a specific source of knowledge of how to start and run a firm as well as a bed for identifying entrepreneurial opportunities. Regions with a high share of service companies have a higher likelihood of attracting new start-ups than regions with a low proportion of service industries (Van Stel, and Storey 2004). One major reason is that the average size of service firms is much smaller, and the employees of those firms are often more exposed to customer contacts, a primary source of entrepreneurial opportunities (Sorensen 2007). Most new firms embody to some extent the norms and practices of the nearby environment (Aldrich, and Ruef 2006) and if there is a large share of service sector in that environment, the probability of a focal individual having worked in, or

knowing entrepreneurs in service sector, increases. As regards the potential differences in regional numbers of start-ups of various types of firms, the question is whether the dynamics of the modern service economy investigated by, among others, Van Stel and Storey (2004) is more or less accentuated in family firms and non-family firms. General samples of family firms tend to be characterized as being less entrepreneurial (Naldi et al. 2007) and frequently more long-term oriented due to their need to maintain relationships with a more diverse set of stakeholders (Sharma, and Manikutty 2005). Family businesses have a stronger tendency to revert to economic structures and practices in place (Trigilia, and Burroni 2009), and compared to these the nonfamily businesses should be more likely to focus their efforts on the emerging opportunities in the service sectors. This leads us to posit that the influence of the service sector in the region will be stronger for non-family businesses: Hypothesis 4b: The size of the service sector will facilitate to a larger extent the emergence of non-family start-ups than family start-ups. Start-ups and Politics Economically, local authorities provide tangible, economic influences that may facilitate or inhibit firm creation, such as the creation or abolition of small firm assistance programs, and through the general structure of local expenditures (Van Stel, and Storey 2004). Further, different local governments can have different attitudes and policies towards firm creation (Garofoli 1994). Sociologically, local authorities wield coercive pressure that can hamper or facilitate the start-up activities of local firms. For instance by indirectly or directly influencing public administrators, application procedures can be delayed or respectively approved rapidly in cases such applications are necessary (DiMaggio, and Powell 1983). While these arguments are built on a top-down logic of how local officials may shape entrepreneurial efforts, the political dominance also reflects to some extent the socio-cultural attitudes of the population towards entrepreneurship and firm creation (Keeble, and Walker 1994). That is, local politics is both a mirror of the local region and

a factor shaping the economic activity in a region. By examining the political influence on firm creation, Davidsson et al. (1994) found out that there is a negative relationship between firm creation and socialist voters in Sweden. However, these voters tended to be in regions with a high share of large manufacturing corporations. Still, no study so far has attended to differences between family businesses and non-family businesses. We believe that the documented correlations between political tendencies and firm creation may be particularly accentuated for family firms, which are less geographically mobile and often have historically stronger ties to a specific region. Following this logic, we set up the following hypothesis: Hypothesis 5: Family start-ups are more likely to appear in politically conservative regions than non-family start-ups. Methodology Data Our theoretical framework posits that the patterns by which family and non-family start-ups emerge differ systematically along a number of regional and economic dimensions. Thus, we describe the prevalence of family start-ups and compare the ecological patterns of these as compared to non-family business start-ups. Furthermore, we analyse in which regions these startups occur and how this is related to the environmental characteristics of the region. The level of analysis in this study is the municipality level. In order to study this, our analysis is based on two longitudinal multi-level databases that cover all regions, companies, and individuals in Sweden between 1991 and 2006 and are provided by Statistics Sweden. The first database, RAMS, provides yearly data on all firms registered in Sweden. Thus, RAMS was used to identify all companies started between 1991 and 2006 in each of the 290 municipalities existing in Sweden. The second database LISA comprises information about all individuals of 16 years of age and older that are registered in Sweden. We have been able to construct linkages across levels of

analyses between the firm-level database RAMS and the individual-level database LISA, as to distinguish between family and non-family businesses by linking individuals to their families. Dependent Variables and Analysis In this study two dependent variables are of interest: the firm births of family businesses and the firm births of non-family businesses. The reason for using the actual number of start-ups instead of the start-up rate (for example measured as the number of new businesses per 1,000 employees) as the dependent variable is that the independent variables using the number of employees as the denominator (for instance proportion of service sector employees or proportion of public sector employees) are also influenced by changes in employment. As a result, these kinds of independent variables could suffer from a pseudo correlation with the start-up rate. Thus, the number of new start-ups seems to be a more appropriate measure (Fritsch, and Falck 2007). In order to obtain the family business start-ups, it is crucial to define what constitutes a family business. It is noteworthy, that in the family business literature no prevalent definition today exists of what constitutes a “family business”. Thus, we define a “family business” as a business where at least two family members are actively engaged in the management and ownership of the business (that is in line with Sonfield, and Lussier 2004; Anderson et al. 2005). In our definition, the term “family” denotes to the nuclear family, that is father, mother, and their children. Thus, given our dataset, an owning and managing family member is someone who obtains his main source of income from being self-employed in the respective company. A family business is a firm where two members of the nuclear family obtain their main source of income from being self-employed in the company. Second, the non-family business start-ups include all companies started by one or more individuals in a respective year and municipality. These companies include all legal forms of companies and appear in all kinds of industries. Since we want to investigate the general patterns of firm births and the potential differences in theoretically derived

predictor variables on the birth of family businesses and non-family businesses, we do not impose certain conditions on the family and non-family business start-ups in terms of size or industry. Independent Variables Regional Population. This variable is important both from a demand-side and a supply-side perspective. From a demand-side perspective, population size sheds light on the current and future size of the market (Davidsson et al. 1994). From the supply side, population size is also important since higher populated areas are likely to produce more entrepreneurs and also possess a higher developed infrastructure than sparsely populated areas (Stuart, and Sorenson 2007). To control for these effects, we enter the population size of each municipality from 1991 until 2006, similarly as prior studies (Braunerhjelm, and Borgman, 2004; Davidsson et al. 1994; Lee et al. 2004). This variable was selected from the official statistical database of Statistics Sweden. Number of Children Born: This predictor variable has rarely been recognized in studies looking at environmental factors encouraging the emergence of start-ups. In our study, short-term regional population growth was highly multicollinear with other important predictor variable, and we therefore decided to use a more long-term indicator of population change in a region, namely the annual number of children born per woman. This variable gives an indication of how the population will develop in the long run, and thus gives some indication of the long-term population growth and prosperity of the municipality investigated. The variable was selected from the official statistical database of Statistics Sweden. Income Per Capita in Municipality. This is an important factor for economic demand of a region since it captures the income distribution (i.e. wealth) across municipalities. Following earlier studies highlighting the importance of income levels (Bull, and Winter 1991; Davidsson et al. 1994; Giannetti, and Simonov 2007), we control for the mean income per capita which

approximates the demand for new goods and services in a local municipality. Also this variable was selected from the official statistical database of Statistics Sweden. Number of Small Businesses. Taking the number of small businesses into account in a regional analysis of start-ups is important for two reasons, one sociological and one economic: First, the existence of a large number of small businesses is supposed to act as role models for other potential entrepreneurs (“small firm incubator hypothesis”). Small businesses also produce employees which are familiar with the operations in small companies and see possible market opportunities, and are thus more likely to start enterprises themselves (Sorensen 2007). This sociological notion highlights that the existence of a high number of small firms in the population generates a favourable socio-economic environment for the start-up process. Second, industrial economic theory suggests that the emergence of new businesses is only possible if no large corporations dominate the market (Comanor, and Wilson 1967). Following Fritsch and Falck (2007) we calculated the number of businesses with fewer than fifty employees in a specific municipality, using the RAMS database (Garofoli 1994; Keeble, and Walker 1994). Urbanization. Urbanization has been considered an important factor in explaining the birth of enterprises. The classification of different degrees of urbanization was provided by Statistics Sweden and comprises nine categories of urbanization ranging from 1 for urban municipality until 9 for other small municipality (1=urban municipality; 2=suburban municipality, 3=large city municipality, 4=medium city municipality, 5=industrial municipality, 6=rural municipality, 7=sparsely populated municipality, 8=other large municipality, 9=other small). Immigrants. The rate of firm formation, and the tendency to engage in family enterprises are known to differ across social and ethnic strata in society. In modern developed economies, the predominant differences concern different ethnicities (Bull, and Winter 1991). We therefore also include a variable denoting to the number of citizens born abroad and having immigrated to

Sweden in each municipality, drawn from Statistics Sweden‟s publicly available database. To investigate the potential of non-linearities, we include both its linear and squared form. Proportion of Service Sector Employees. A substantial share of new businesses tend to be set up in the service industry (Braunerhjelm, and Borgman 2004; Fritsch, and Falck 2007) which has lower entry barriers than for instance the industry sector (Van Stel, and Storey 2004). We control for the share of individuals employed in service industry in relation to the municipality‟s population using two digit level of the Swedish Standard Industrial Classification served as a basis for determining the service industries (SNI 50-74), obtained from the LISA database. Politics: We include a variable indicating the political dominance in each municipality since this can imply different supportive environments with regard to the creation of new firms. To distinguish between the effects of local politics on the emergence of family businesses and nonfamily businesses, we define our measure of local politics as a time-variant variable taking the value -1 for socialistic majority, 1 for right-wing majority, and 0 for a mixed (coalition) majority. This was taken from Statistics Sweden‟s public databases. Control Variables Proportion of Public Sector Employees. Some research suggests that a high proportion of public private sector employees have a negative impact on the firm creation since a high number of secure employment positions might discourage entrepreneurial activity (Littunen. and Hyrsky 2000; Özcan, and Reichstein 2009). However, studies to date disagree on the importance and direction of these effects (Davidsson et al. 1994. We thus control for this effect by measuring the relative ratio of employees in the public sector in proportion to the municipality‟s population (Davidsson et al. 1994; Gianetti, and Simonov 2007). The variable was calculated through the LISA database by using two-digit level of the Swedish Industrial Classification (SNI 75-99). All predictor and control variables were lagged one year to avoid simultaneity bias.

Analytical Procedures In order to draw causal inferences from our theoretically derived environmental-level variables and the number of start-ups of family and non-family businesses, we estimate regression models to test for differences in the theoretically derived predictor variables across the models. Since the number of started family and non-family businesses take on non-negative integers (for example 0 ,1 , 2, ...) we employ count data analysis. The variance of the outcome variable exceeds the mean, meaning that the data is „overdispersed‟ which violates the assumptions of the poisson model. We therefore use the negative binomial model which allows for the specific parametric assumptions by including a parameter, α, indicating how the variance differs from the mean (Cameron, and Trivedi 1998). All variables, their modal values, and the correlations are displayed in Table 1.1 -----------------------------------------Insert Tables 1 and Table 2 here -----------------------------------------Results Table 2 shows negative binomial regression models of firm births across all Swedish municipalities during the time period of analysis. We show separate models for family start-ups and non-family start-ups, the latter by far representing the majority of firm births. To test our hypotheses and formally investigate whether the independent variables affect the emergence of 1

The table indicates low to moderate correlation between the variables with the exception of the correlations between the dummy for urbanity and the two variables denoting the number of residents with a foreign background (=0.797 for the linear and 0.692 for the squared term) and the squared proportion of residents with a foreign background and the number of small businesses in a municipality (0.720). Variance Inflation Factors (VIF) for all variables were below the generally accepted threshold of 10 except for the two variables population size and number of small businesses amounting to 23.92 and 19.17, respectively. Multicollinearity is common in studies of organizational births and commonly means that although coefficients will be unbiased standard errors will be inflated, rendering hypotheses testing more conservative (Cattani et al. 2003). We therefore transformed the variables population size and number of small businesses into their logarithmic versions. We also estimated 500 bootstrapped models based on a random 90 percent sample as well as estimating models omitting the top and bottom 5 percentiles of municipalyear start-ups. This did not significantly alter the results, which therefore seem to be robust to multicollinearity.

family start-ups and non-family start-ups differentially, we used the seemingly unrelated estimation procedure (STATA command suest, see Weesie 1999) that tests whether the effect of a predictor x on a binary outcome y is the same as the effect on another binary outcome. Looking first at population size, we find that the influence of population size is positive for both non family start-ups (0.316, p<0.01) and family start-ups (0.224, p<0.10), although the latter is significant only on the 10 percent level. Further, the suest test statistic (1.21) reveals that there is no significant difference between non-family and family start-ups. This means that although the positive effect of population size is statistically more ascertained for non-family start-ups, we cannot ascertain there is a systematic difference in how the size of the local market affects nonfamily start-ups and family start-ups, meaning that we cannot accept hypothesis 1a. As regards the impact of number of children born per woman on the emergence of start-ups, we find that the number of children born is positively associated with non-family start-ups (0.152, p<0.01) but negatively associated with family start-ups (-0.246, p<0.01). The suest statistic (186.80, p<0.01) affirms that there is a significant difference between the number of start-ups of family and non-family businesses across the municipalities of investigation. This supports hypothesis 1b. Regarding the impact of regional wealth on the emergence of start-ups, the coefficient for income per capita in municipalities vindicate that family start-ups differ significantly from non-family start-ups (suest statistic: 202.19, p < 0.01). The level of income has a positive effect on non-family start-ups (0.967, p <0.01) while this influence is negative (-6.909, p<0.01) for family start-ups. This supports hypothesis 1c which stated that the level of income will be of higher importance for non-family start-ups than for family start-ups. Considering the effect of rural municipalities on firm birth, the suest test statistic (3.59, p< 0.05) show that nonfamily start-ups significantly differ from family-start-ups. The results reveal that while family start-ups are positively associated with rural municipalities (0.094, p<0.10) we observe a

negative, albeit not significant (-0.014) relationship for non-family start-ups. The corresponding dummy for urban region is negative and not significant for both types of start-ups, and the suest statistic reveal there is no significant difference between these two groups. Thus, we find support for hypothesis 2. Turning next to the effect of immigrants on family start-ups, we see that a significant difference between family and non-family start-ups exists (suest statistics: 67.30, p<0.01 for immigrants; 69.48, p<0.01 for immigrants²). The coefficients show a weak positive relationship between the proportion of immigrants in a municipality and the number family start-ups (0.000, p<0.01). However, the squared coefficient shows that as the proportion of immigrants increases to a certain level, this effect becomes negative (-0.008, p<0.001). The opposite effect can be observed for non-family start-ups, implying that a high proportion of immigrants has a slight negative influence (-0.000, p<0.05) but that a very high proportion of immigrants has a positive effect (0.003, p<0.0001). Since hypothesis 3 stated that family start-ups are more prone to appear in regions with a high number of immigrants, we find only partial support in that family start-ups seem more prone to appear in regions with a medium-to high number of immigrants, but are less prone to appear in regions with very high number of immigrants. A central point of our theoretical framework is that family firms are less directly economically affected and more strongly regionally embedded than non-family firms. One way of investigating this is to test if the availability of role models affect the emergence of family and non-family start-ups in differential ways. We test this by examining the coefficient for the variable number of small businesses in the municipality (our proxy for the availability of role models), observing that the suest statistics (12.44, p<0.01) reveal a significant difference between family and non-family start-ups, with an apparently stronger effect on family start-ups (0.876, p<0.01) compared to nonfamily start-ups (0.609, p<0.01). This supports hypothesis 4a. Looking next at the influence of

the service sector on regional start-ups, we see that there exists a significant difference between family and non-family start-ups (suest statistics: 30.30, p< 0.01). Furthermore, the service sector exerts a positive relationship (0.856, p<0.01) on non-family start-ups while it has a negative influence on family start-ups (-0.706, p<0.01). Thus, we can find strong evidence for our hypothesis 4b. Finally, the coefficient for local political majority reveal a positive relationship between right-wing politics and family start-ups (0.058, p<0.01) providing evidence that family start-ups are more likely to emerge in municipalities with a non-socialistic government. No significant relationship can be found for non-family start-ups (0.001, p>0.10). The suest statistic (13.55, p<0.01) confirms that there is a significant difference between the two dependent variables and politics, leading us to confirm also hypothesis 5. Discussion In this study we set out to investigate the regional bed for family start-ups and non-family start-ups. Using a rich multi-level data set from 1991 to 2006, we found that while economic factors strongly affect the number of non-family start-ups, the number of family start-ups was more strongly tied to non-economic factors. Specifically, we found that regional wealth, size of the service sector, and long-term population growth in a municipality positively affected the number non-family start-ups, as we hypothesized. Conversely, the number of family start-ups was positively associated with rural municipalities, the number of other small businesses in the municipality, and the municipality being governed by right-wing politics, also as hypothesized. As regards the relationship between level of immigrants and start-ups, our findings were somewhat more complex than anticipated. The analyses indicate that the association between immigrants and non-family start-ups is weakly U-shaped, with a high proportion of immigrants having a slight negative influence on the number of non-family start-ups but that a very high proportion of immigrants having a positive effect. For family start-ups, our analyses indicate an

inverse U-shaped relationship between the proportion of immigrants in a focal municipality and the number family start-ups. Although this association is only partly as we hypothesized, we think the relationship between immigration and the composition of entrepreneurial activity in varying regions across developed nations is an issue that deserves more attention. The increasing globalization of economic activity and the increased level of migration of both high-skilled and low-skilled individuals suggests that the issue of how migration affects economic activities in general, and entrepreneurial processes such as the rate of firm formation or firm growth in particular, has only begun to be addressed. Our paper represents an early attempt in this respect, and hope that future studies can develop, challenge, or replicate our findings also in other empirical contexts. Taken together, the findings in this paper indicate that family start-ups are more susceptible to the local non-economic context than non-family start-ups. These are imperative issues for both research and public policy. For research in entrepreneurship, our paper indicates that „the geographic connection‟ is an important yet undertheorized area affecting the probability of start-ups as well as what type of start-ups that tend to emerge in different regions. Our paper furthermore highlights the importance of distinguishing between various types of startups – family businesses and non-family businesses – in aggregate analyses of entrepreneurial processes and their potential outcomes. Our focus on how micro-level processes are affected by meso-level ecological characteristics lay the ground for further in-depth studies about the nexus between entrepreneurship and family business, potentially looking not only at the start-up of such businesses but also how they may grow, innovate or eventually cease to exist (Hellerstedt et al. 2010). For research in economic geography, our paper hence offers an empirical substantiation of how entrepreneurial families constitute important influences on the entrepreneurial processes, that is the start-up of companies.

References Aldrich, H. E., and M. Ruef (2006). Organizations Evolving. London: SAGE Publications Ltd. Aldrich, H. E., and J. E. Cliff (2003). “The Pervasive Effects of Family on Entrepreneurship: Toward a Family Embeddedness Perspective,” Journal of Business Venturing, 18 (5), 573596. Aldrich, H. E. (1990). “Using an Ecological Perspective to Study Organizational Founding Rates,” Entrepreneurship Theory and Practice, 14 (3), 7-24. Anderson, A. R., S. L. Jack, and S. Drakopoulou Dodd (2005). ”The Role of Family Members in Entrepreneurial Networks: Beyond the Boundaries of the Family Firm,” Family Business Review, 18 (2), 135- 154. Birley, S. (1986). “The Role of New Firms: Births, Deaths and Job Generation,” Strategic Management Journal, 70 (4), 361-376. Braunerhjelm, P., and B. Borgman (2004). “Geographical Concentration, Entrepreneurship and Regional Growth: Evidence from Regional Data in Sweden, 1975–99,” Regional Studies, 38 (8), 929-947. Brockhaus, R. H. (1994). “Entrepreneurship and Family Business Research: Comparisons, Critique and Lessons,” Entrepreneurship Theory and Practice, 19 (11), 25-38. Bull, I., and F. Winter (1991). “Community Differences in Business Births and Business Growths,” Journal of Business Venturing, 6 (1), 29-43. Cameron, C. A., and P. K. Trivedi, (1998). Regression Analysis of Count Data. Cambridge: Cambridge University Press. Cattani, G., J. M. Pennings, and C. M. Wenzel (2003). “Spatial and Temporal Heterogeneity in Founding Patterns.” Organization Science, 14(6), 670-685. Coffey, W., and M. Polèse (1985). “Local Development: Conceptual Bases and Policy Implications,” Regional Studie, 19 (2), 85-93. Colli, A., P. Fernández Pérez, and M. B. Rose (2003). “National Determinants of Family Firm Development? Family Firms in Britain, Spain, and Italy in the Nineteenth and Twentieth Centuries,” Enterprise & Society, 4, 28-64. Comanor, W. S., and T. A. Wilson (1967). “Advertising Market Structure and Performance,” Review of Economics and Statistics, 44 (4), 423–440. Dahl, M. S., and O. Sorenson (2009). “The Embedded Entrepreneur”, European Management Review, 6, 172-181. Davidsson, P., L. Lindmark, and C. Olofsson (1994). “New Firm Formation and Regional Development in Sweden,” Regional Studies, 28 (4), 395-410. Davidsson, P., and J. Wiklund (2001). “Levels of Analysis in Entrepreneurship Research: Current Research Practice and Suggestions for the Future,” Entrepreneurship, Theory and Practice, 25(4), 81-100. DiMaggio, P. J., and P .W. Powell (1983). The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review, 48 (2), 147-160. Dobrev, S. (2005). “Career Mobility and Job Locking,” Social Science Research, 34 (4), 800820. Donckels, R., and E. Fröhlich (1991). “Are Family Businesses Really Different? European Experiences from STRATOS,” Family Business Review, 4 (2), 149-160. Dyer, G. W., and W. Handler (1994). “Entrepreneurship and Family Business: Exploring the Connections,” Entrepreneurship Theory and Practice, 19 (1), 71-83.

Feldman, M. P., J. Francis, and J. Bercovitz (2005). “Creating a Cluster while Building a Firm: Entrepreneurs and the Formation of Industrial Clusters,” Regional Studies, 39(1), 129-141. Fritsch, M., and O. Falck (2007). “New Business Formation by Industry over Space and Time. A Multidimensional Analysis,” Regional Studies, 41 (2), 157-172. Garofoli, G. (1994). “New Firm Formation and Regional Development. The Italian Case. Regional Studies,” 28(4), 381-393. Gartner, W. B. (1988). ““Who is an Entrepreneur?” Is the Wrong Question,” American Journal of Small Business, 12 (4), 11–32. Gianetti, M., and A. Simonov (2009). “Social Interactions and Entrepreneurial Activity,” Journal of Economics and Management Strategy, 18 (3), 665-709. Gimeno, J., T. Folta, A. Cooper, and C. Woo (1997). “Survival of the Fittest?:Entrepreneurial Human Capital and the Persistence of Underperforming Firms,” Administrative Science Quarterly, 42 (4), 750-778. Hannan, M. T., G.R. Carroll, E. A. Dundon, and J. C. Torres (1995). “Organizational Evolution in a Multinational Context: Entries of Automobile Manufacturers in Belgium, Britain, France Germany and Italy,” American Sociological Review, 60(4), 509-528. Hammarstedt, M. (2004). “Self-Employment Among Immigrants in Sweden – An Analysis of Intragroup Differences,” Small Business Economics, 23 (2), 115-126. Hellerstedt, K., M. Nordqvist, and K. Wennberg (2010). “Performance Effects of Ownership Transitions in Family firms,” in Frontiers of Entrepreneurship Research. Ed. A. Zacharakis, S. Alvarez, M. Cardon, J. De Castro, and F. Delmar. College Park: Babson College. Johannisson, B., M. Ramirez-Pasillasm, and G. Karlsson (2002). “The Institutional Embeddedness of Local Inter-Firm Networks: A Leverage for Business Creation,” Entrepreneurship & Regional Development, 14 (4), 297-315. Johnson, P., and S. Parker (1996). “Spatial Variations in the Determinants and Effects of Firm Births and Deaths,” Regional Studies, 30 (7), 679-688. Keeble, D., and S. Walker (1994). New Firms, Small Firms and Dead Firms: Spatial Patterns and Determinants in the United Kingdom. Regional Studies, 28(4), 411-427. Lee, S. Y., R. Florida, and Z. J. Acs (2004). “Creativity and Entrepreneurship: A Regional Analysis of New Firm Formation,” Regional Studies, 38 (8), 879-891. Littunen, H., and K. Hyrsky (2000). “The Early Entrepreneurial Stage in Finnish Family and Nonfamily Firms,” Family Business Review, 13 (1), 41-54. Low, M. B., and I. C. MacMillan (1988). “Entrepreneurship: Past Research and Future Challenges,” Journal of Management, 14 (2), 139-161. Malecki, E. (1993). “Entrepreneurship in Regional and Local Development,” International Regional Science Review, 16 (1-2), 119-153. Mason, C. M., and R.T. Harisson (2006). “After the Exit: Acquisitions, Entrepreneurial Recycling and Regional Economic Development,” Regional Studies, 40(1), 55–73. Naldi, L., M. Nordqvist, K. Sjöberg, and J. Wiklund (2007). “Entrepreneurial Orientation, Risk Taking, and Performance in Family Firms,” Family Business Review, 20 (1), 33-47. Nordqvist, M., and L. Melin (2010). “Entrepreneurial Families and Family Firms,” Entrepreneurship and Regional Development, 22 (3-4), 211-239. Özcan, S., and T. Reichstein (2009). “Transition to Entrepreneurship from the Public Sector: Predispositional and Contextual Effects,” Management Science, 55, 604–618. Pe‟er, A., and I. Vertinsky (2008). “Firm Exits as a Determinant of New Entry: is there Evidence of Local Creative Destruction?,” Journal of Business Venturing, 23(3), 280–306.

Pe'er, A., I. Vertinsky, and A. King (2006). “Who Enters, Where and Why? The Influence of Capabilities and Initial Resource Endowments on the Location Choices of De Novo Enterprises.” Strategic Organization, 6, 119-149. Pennings, J. M (1982). “Organizational Birth Frequencies: An Empirical Investigation,” Administrative Science Quarterly, 25 (1), 120-144. Reynolds, P. (1994). “Autonomous Firm Dynamics and Economic Growth in the United States 1986-1990.” Regional Studies, 28 (4), 429-442. Romanelli, E. (1991). “The Evolution of New Organizational Forms,” Annual Review of Sociology, 17(1), 79-103. Ruef, M. (2000). “The Emergence of Organizational Forms: A Community Ecology Approach,” American Journal of Sociology, 106 (3), 658-714. Sanders, J. M, and V. Nee (1996). “Immigrant Self-Employment: The Family as Social Capital and the Value of Human Capital,” American Sociological Review, 61 (2), 231-249. Sharma, P., and S. Manikutty (2005). “Strategic Divestments in Family Firms: Role of Family Structure and Community Culture,” Entrepreneurship Theory and Practice, 29 (3), 293311. Singh, J. V., and C. J. Lumsden (1990). “Theory and Research in Organizational Ecology,” Annual Review of Sociology, 16, 161-195. Sonfield, M.C., and R. N. Lussier, (2004). First-, Second-, and Third-Generation Family Firms: A Comparison. Family Business Review, 17 (3), 189- 202. Sorensen, J. (2007). “Bureaucracy and Entrepreneurship: Workplace Effects on Entrepreneurial Entry,” Administrative Science Quarterly, 52, 387-412. Stiles, C. H., and C. S. Galbraith (2003). Ethnic Entrepreneurship: Structure and Process. Oxford: Elsevier. Stuart, S., and O. Sorenson (2003). “The Geography of Opportunity: Spatial Heterogeneity in Founding Rates and the Performance of Biotechnology Firms,” Research Policy, 32, 229253. Trigilia, C., and L. Burroni (2009). “Italy: Rise, Decline and Restructuring of a Regionalized Capitalism,” Economy and Society. 38 (4), 630-653. Van Stel, A. J., and D. Storey (2004). “The Link between Firm Births and Job Creation. Is there a Upas Tree Effect,” Regional Studies, 38 (8), 893-909. Weesie, J. (1999). “Seemingly unrelated Estimation and the Cluster-Adjusted Sandwich Estimator,” Stata Technical Bulletin, 52, 34-47. Zahra, S. A., J. C. Hayton, and C. Salvato (2004). “Entrepreneurship in Family vs. Non-Family Firms: A Resource-Based Analysis of the Effect of Organizational Culture.” Entrepreneurship, Theory and Practice, 28(4), 363-382.

Table 1 Variable Modal Values and Correlation Matrix

Mean 1 Family start-ups

St. Dev.

Min

Max

1

2

3

4

5

6

7

8

9

10

11

95.23

113.11

1

1712

288.11

669.56

16

12220

0.796

9.24

2.49

0.00

13.59

0.341

0.178

0.00

0.03

-0.05

0.18

0.046

0.164

0.216

6.59

0.99

0.00

10.95

0.767

0.619

0.394

0.236

0.00

4938.44

-1722

73465

0.768

0.896

0.189

0.106

0.467

31.50

26.95

0.00

274.20

0.783

0.840

0.298

0.181

0.720

0.813

1.92

0.32

0.91

3.18

-0.233 -0.137

-0.374

-0.175

-0.230

-0.116

-0.210

0.27

0.09

0.00

0.73

0.420

0.294

0.368

0.623

0.360

0.550

-0.220

0.38

0.08

0.00

0.70

-0.009 -0.031

0.186

-0.009

0.178

-0.054

-0.033

-0.024

0.047

11 Politics

0.03

0.88

-1.00

1.00

0.047

0.035

-0.085

0.072

0.085

-0.022

-0.017

0.217

0.146

-0.059

12 Urban municipality

0.01

0.10

0.00

1.00

0.571

0.800

0.125

0.048

0.340

0.797

0.692

-0.099

0.250

-0.064

-0.021

13 Rural municipality

0.10

0.30

0.00

1.00

-0.042 -0.075

-0.071

-0.146

-0.092

-0.080

-0.181

0.094

-0.133

-0.049

0.222

2 Non-family start-ups 3 (ln)Population Size Income per Capita (mean centered) (ln)Number of Small 5 businesses Immigrants (mean 6 centered) 4

7 Immigrants² 8 Children born in region Proportion Service Sector 9 Employees Proportion Public Sector 10 Employees

All correlations above +/- 0.03 are significant at a 5 % level

0.425

12

-0.035

Table 2 Seemingly Unrelated (SUR) Regression Models of the Negative Binomial Type on Family Start-ups and Non-Family Start-Ups, Respectively

(ln) Population Size Children per women Income per Capita (ln)Number of Small businesses Immigrants Immigrants ² Proportion of Service Sector Employees

Non-family Start-ups 0.316*** (0.071) 0.152*** (0.011) 0.967*** (0.137)

Family Start-ups 0.224+ (0.115) -0.246*** (0.026) -6.909*** (0.536)

Chow test of significant differences 1.21 H1a: rejected 186.80*** H1b: affirmed 202.19*** H1c: affirmed

0.609*** (0.070) -0.000* (0.000) 0.003*** (0.001)

0.876*** (0.113) 0.000*** (0.000) -0.008*** (0.001)

12.44***

0.856***

-0.706**

30.30***

(0.250)

H4b: affirmed

-0.373

14.19***

(0.100) Proportion of Public Sector 0.571*** Employees (0.086) Politics 0.001 (0.008) Urban Municipality -0.277*** (0.066) Rural Municipality -0.014 (0.027) Constant -2.836*** (0.257) Log Likelihood -23806.589 Pseudo-R2 (McFadden's) 0.2701 AIC Value 47651.179 Municipality-year Obs. 4,906 Number of Municipalities: 290

(0.231) 0.058*** (0.016) -0.500** (0.163) 0.094+ (0.052) -2.699*** (0.446) -22270.128 0.1822 44578.256 4,906 290

67.30*** H3: affirmed 69.48*** H3: rejected

13.55** H5: affirmed 1.80 3.59* H2: affirmed

Note: Year dummies and 7 additional regional dummies included but not reported. Standard errors clustered by municipality in parentheses. + p<0.10 * p<0.05; ** p<0.01; *** p<.001; (two-tailed).