OWNERSHIP SOCIAL CAPITAL IN THE PRIVATELY-HELD FIRM: A STRUCTURAL MODEL WITH MODERATOR EFFECTS OF OWNERMANAGEMENT OVERLAP Marta M. Berent-Braun Center for Entrepreneurship Nyenrode Business Universiteit Lorraine M. Uhlaner, corresponding author EDHEC Business School 24 avenue Gustave Delory 59057 Roubaix, France +33 320154449 [email protected] Roberto Flören Center for Entrepreneurship Nyenrode Business Universiteit ACADEMIC ABSTRACT Based on a dataset of 560 Dutch privately held firms, this study first presents a measurement model, together with tests of convergent and discriminant validity of three separate factors (structural, relational, and cognitive) for ownership social capital, as represented by family involvement, quality of relationships amongst owners, and shared vision, respectively. Furthermore, a structural equation model using maximum likelihood estimation suggests that while the three factors together enhance prediction of financial performance, structural models which take moderation effects of ownership-management overlap are better fitting. Furthermore, models including mediating effects amongst social capital variables eliminate nonsignificant paths. Key words: ownership group, social capital theory, family involvement, financial performance, owner-management overlap

EXECUTIVE SUMMARY Thesis This study tests whether or not three aspects of ownership social capital, representing cognitive, relational and structural dimensions (shared vision, quality of relationships between owners, and family involvement, respectively) influence financial performance of the business. Methodology Survey data collected from the director as key informant are based on a stratified random sample of 560 Dutch private firms. Selected firms had between two and twenty owners. Confirmatory factor analysis was used to confirm a three factor solution for social capital. Structural Equation Modeling with a maximum likelihood estimation was used to analyze proposed relationships. Findings A baseline model of ownership social capital enhances prediction of financial performance by control variables (including company age, company size, number of owners, sector dummies) by 5.5%. Moderation effects of ownership-management overlap improve both the model’s fit and its predictive power. Where ownership and management overlaps completely, shared vision has a direct effect, mediating the relationships of family involvement and quality of relationships with financial performance. By contrast, in conditions of partial to nonoverlap, family involvement and quality of relationships directly effect financial performance, the latter mediating the effect of shared vision. Implications for theory and practice This study uses concepts from social capital and group dynamics to develop and test a model of ownership social capital in the private firm. The findings suggest that shared goals and objectives among owners and other qualities of a good team (trust, cooperation) significantly influence a firm’s financial performance. How the findings can be implemented Introduction of teambuilding activities for owners may be an effective tool to enhance company performance. Degree of overlap between ownership and management may determine which dimensions of social capital are most critical and thus what type of training program should be designed and implemented.

INTRODUCTION This study applies social capital theory (Adler and Kwon, 2002; Nahapiet and Ghoshal, 1998; Leana and Van Buren, 1999) to the owning group in private firms. More specifically, we explore the relationship between certain ownership characteristics and financial performance. While initially used for community studies (e.g. Loury, 1977), the application of social capital to the organizational level of analysis is well established (e.g. Nahapiet and Ghoshal, 1998; Leana and Van Buren, 1999; Adler and Kwon, 2002; Lee, 2009). However, whereas past organization-level research tends either to focus on employees (e.g. Nahapiet and Ghoshal, 1998; Leana and Van Buren, 1999) or family (e.g. Arregle et al., 2007; Carr et al., 2011; Mustakallio et al., 2002), such research typically ignores the owning group, especially among nonfamily-owned firms. Furthermore, while there is a growing body of research on ownership, it focuses primarily on structural aspects, including ownership concentration (e.g. Thomsen and Pedersen, 2000; Wu et al., 2007), proportion of family owners (Anderson and Reeb, 2003; Dyer, 2006; O’Boyle et al., 2012), proportion of institutional, private, government ownership (Pedersen and Thomsen, 2003), and the number of owners (Madsen et al., 2007). Though generally ignored as a distinct social group, especially in the organizational behavior literature, business owners can be viewed as a social group distinguished from other social entities in the firm by possession of ownership rights in the company (BerentBraun and Uhlaner, 2012). To highlight our emphasis not only on structural but also relational and cognitive aspects of the owning group, we use the term, ownership social capital, to represent “the character of social relations” amongst owners. The potential for owners as a source of social capital was introduced first in the family business literature (see Mustakallio et al., 2002) and research on owner commitment (Uhlaner et al., 2007). Ownership can be considered from a narrow legal angle or “from a broader perspective that encompasses knowledge, skill, integrity, responsibility, trust, and mutual consideration” (Williams, 1992, p. 161). In the context of the family business literature, these attributes capture a “spirit of ownership” that facilitates family business prosperity by building a “close-knit, effective team” of family owners (Williams, 1992, p. 161). Although the “spirit of ownership” was used by Williams (1992) with respect to family owners, based on the trust and group dynamics literature, such attributes as integrity, responsibility, trust, or mutual consideration can be seen as beneficial for any group of owners, regardless of family linkages. The key research question in the present study is as follows: “How does ownership social capital affect business performance?” We examine how the three aspects of ownership social capital, including shared vision, the quality of relationships between owners, and family involvement influence financial performance of the business. In addition to firm characteristics we control for the number of individual owners. For each of the hypothesized main effects we further test for the possible moderating effect of overlap between ownership and management. This paper makes two important contributions to the literature on social capital. First of all, it enhances our understanding of the role of owners in the privately-held business by testing empirically how the three components (structural, relational and cognitive) of ownership social capital influence financial performance of the firm. In spite of a growing

interest in family social capital, to our knowledge studies on ownership social capital are still relatively scarce. Secondly, by testing for moderating effects of ownershipmanagement overlap, and including family involvement as an independent variable, we are able to separate more clearly than in past research whether social capital effects are due to family, ownership or management. The remainder of the paper is structured as follows. First, we present a framework and rationale for hypotheses, including an overview of the relevant literature on social capital, after which we present arguments for predicted relationships between ownership social capital and financial performance of the business. The following two sections present the method and results followed by the discussion section, which covers the theoretical and practical implications of the results, limitations of the current study, and recommendations for future research. FRAMEWORK AND RATIONALE Ownership social capital and business performance Social capital developed within organizations can enhance business performance in a variety of ways (Bolino et al., 2002; Lee, 2009). It may lower transaction costs resulting from unfair or opportunistic behavior (Lee, 2009; Walker et al., 1997); enhance information flow (Walker et al., 1997; Wu, 2008); and build intellectual capital (Leana and Van Buren, 1999; Nahapiet and Ghoshal, 1998). It may also facilitate coordination of efforts (Lazega and Pattison, 2001; Lin, 2001) and facilitate collective work (Putnam, 1993b). Thus, social capital can be regarded as a significant source of competitive advantage for a firm (Lee, 2009). As Bolino et al. (2002, p. 507) note, “previous research indicates that social capital is an important resource because individuals work together more effectively when they know one another, understand one another, and trust and identify with one another”. Concurring findings from a recent meta-analysis by Westlund and Adam (2010) conclude that social capital is linked to a firm’s financial performance. The overall framework tested in this study is presented in Figure 1. In the remainder of this section, we elaborate on the rationale and present each hypothesis in turn. Shared vision (cognitive ownership social capital) and financial performance We propose, first of all, that firms in which the ownership group is unified by a shared vision will achieve better performance than those with a less unified group of owners. The concept of shared vision draws, first of all, on Nahapiet and Ghoshal’s cognitive dimension of social capital, which they define as “shared representations, interpretations, and systems of meaning among parties” (Nahapiet and Ghoshal, 1998, p. 244). However, it also draws on Leana and Van Buren’s concept of associability, which the latter identify as an aspect of organizational social capital and define as the “willingness and ability of participants in an organization to subordinate individual goals and associated actions to collective goals and actions” (Leana and Van Buren, 1999, p. 541). This willingness to think in a way which benefits the group as a whole, rather than the individual alone, can facilitate consensus amongst members about the key objectives of the business. Past empirical research in the field of group dynamics finds that group consensus about goals is positively associated with firm performance (Dess, 1987; Guth and MacMillan, 1986). When business goals are agreed upon, a stable and clear strategy can be developed which, in turn, facilitates the

work of management. A common agreement on goals, furthermore, limits disputes regarding business operations and strategy. A similar argument can be made for the owning group. Without a shared vision, the owners may pull the business in opposite directions since the firm may serve different goals for different owners. For instance, an owner may demand higher dividends or even pull his or her investment out of the business, motivated by the desire to receive higher rents elsewhere. Anecdotal evidence supports the view that many firms stop due to a failure amongst owners to agree about strategy or objectives. On the contrary, a group of owners unified by a shared vision will spend less effort to manage relational conflicts and more on actual business issues (Ensley and Pearson, 2005). A shared understanding amongst owners also enhances business performance by reducing unexpected behaviors (Weick, 1995) and thus decreasing uncertainty within the firm. An ownership group that speaks with one voice will make better quality decisions (Mustakallio et al., 2002), which can be especially important in an unstable business environment. Other empirical research, based on Dutch private firms, concludes that collective norms and goals as well as owner commitment are positively associated with financial performance (Uhlaner et al., 2007). Taken together, these arguments lead to the first hypothesis as follows: Hypothesis 1: The greater the shared vision amongst members of the ownership group, the better the financial performance of the business will be. FIGURE 1 Theoretical Model Cognitive Social Capital Shared Vision

Ownership-Management Overlap

H1 Relational Social Capital Quality of Relationships

Structural Social Capital Family Involvement OWNERSHIP SOCIAL CAPITAL

H4 a, b, c

Financial Performance of the Business H2

H3

Control variables: Age Size Sector Number of owners

Quality of relationships (relational ownership social capital) and financial performance Second, we propose that the quality of relationships between owners, which includes level of trust, honesty, cooperation, and teamwork, has a positive effect on firm performance. Although empirical research on the quality of relationships between owners and its

influence of business performance is scarce, a study by Bosma et al. (2004) shows that social capital of the business founder, as represented by the emotional support of a spouse, increases the chance of survival and profitability of Dutch start-ups. It is also a commonly shared view in the social capital literature that good relationships, such as trust and reliability, may improve organizational functioning (Bouty, 2000; Jones and George, 1998). Workgroups in which members have good relationships “may be more flexible, better able to adapt to a changing environment and [are] higher performing” (Bolino et al., 2002, p. 510). Furthermore, in the teams literature, constructive interpersonal relationships are characteristics of a team (Huszczo, 1996). In this stream of literature, there is a large body of evidence that good interpersonal relationships and teamwork enhance performance of the business (Cohen and Bailey, 1997; Combs et al., 2006; Guthrie, 2001). Trust and honesty is seen in the organizational behavior literature as an organizing principle (McEvily et al., 2003) and is a better coordination and monitoring mechanism than those prescribed by agency theory (Mustakallio et al., 2002). Thus, organizational performance is enhanced as owners who trust each other and are honest with each other will incur less transactional costs. Summarizing, this feeling of trust and honesty, combined with a good cooperative relationship between owners make an ownership group function as a team. When owners work as a team, rather than as separate individuals, the business benefits since decisions are made faster and there is less conflict detrimental to business performance. Thus, the second hypothesis is formulated as follows: Hypothesis 2: The better the quality of relationships between members of the ownership group, the better the financial performance of the business will be. Family involvement (structural ownership social capital) and financial performance Our next hypothesis proposes that the family involvement in the business positively influences firm performance. There is a vast body of literature that argues that family involvement has an influence on business performance, although the direction of this relationship is not clear and can be argued both as a positive and negative relationship (see for instance Dyer, 2006; Habbershon et al., 2003; O’Boyle et al., 2012). Habbershon and Williams (1999) provide an overview of the factors why family firms may perform differently than their nonfamily counterparts, including a working environment enhancing employee care and loyalty, a higher work efficiency of family members, the reduction of transactional and agency costs, or efficient decision making. Arregle et al. (2007) argue for advantages of organizational social capital in the family firm versus the nonfamily firm based upon differences in stability, social interaction, interdependence and closure (Arregle et al., 2007; Nahapiet and Ghoshal, 1998). Stability enhances organizational social capital by facilitating cooperation and creating trust (Hitt et al., 2002; Leana and Van Buren, 1999; Putnam, 1993a). The ownership group in a family business typically is more stable than in the nonfamily firm (Arregle et al., 2007). Firstly, a family – as a social unit – changes very slowly. Social interaction is a necessary condition for social capital to emerge (Bourdieu, 1986). Social interaction in the ownership group of a family firm is more intense than that in the ownership group of a nonfamily firm (Arregle et al., 2007). Family owners have the possibility of frequent contact, even beyond informal family gatherings. Interdependence enhances social capital by maintaining expectations and obligations (Nahapiet and

Ghoshal, 1998). Arregle et al. (2007) argue that owners in family firms are more interdependent than owners in nonfamily firms since the firm represents a shared heritage and is typically the main family asset. Also exit options for owners in family firms are often limited due to the difficulty with price evaluation of shares, or even possible ostracism by other family owners (Arregle et al., 2007). Finally, social capital depends upon the closure of social relationships that is, the presence of clear social boundaries separating a social unit from other units (Etzioni, 1996). According to research by social psychologists, closure facilitates norms, trust, and identity (Coleman, 1990). The ownership group in a family business is characterized by stronger closure than their nonfamily counterparts since the group boundaries are designated though a strong relation of kinship (Arregle et al., 2007). In a recent meta-analysis of the family involvement and firm performance relationship, O’Boyle et al. (2012) conclude that there is no relationship between family involvement and financial performance. However, to be consistent with the social capital arguments, we state the third hypothesis in the affirmative as follows: Hypothesis 3: The higher the family involvement in the business, the better the financial performance of the business will be. The overlap between management and ownership as a moderating variable The purpose of the final set of hypotheses is to tease apart ownership versus management effects. According to the governance literature, the overlap between owners (principals) and managers (agents) can influence performance of the business. It is further assumed that under the condition of complete overlap between ownership and management (even with two or more owners), goals are more likely to be aligned between management and ownership and therefore monitoring mechanisms are not needed to oversee owners’ interests (Jensen and Meckling, 1976). Therefore, the relationships proposed in the model are compared for two conditions: 1) partial to no overlap between ownership and management; and 2) complete overlap. Although the rationale should be similar for companies where ownership and management overlap completely, and where they do not, in the condition of 100% overlap, one would not necessarily expect a savings in agency monitoring costs, since owners and managers represent the same people. In reviewing each of the hypotheses, we therefore only predict positive ownership effects on financial performance in the case where ownership and management are at least partially nonoverlapping. We state the following hypotheses as follows: Hypothesis 4a: Under the condition of partial to no overlap between owners and managers, shared vision amongst members of the ownership group has a positive and direct effect on financial performance of the business. Hypothesis 4b: Under the condition of partial to no overlap between owners and managers, the quality of relationships between members of the ownership group has a positive and direct effect on financial performance of the business . Hypothesis 4c: Under the condition of partial to no overlap between owners and managers, family involvement in the business has a positive and direct effect on financial performance of the business.

Control variables In order adequately to assess the relationships proposed in the hypotheses presented above we control the sample for a range of variables. Company size, age and sector are generally cited as important control variables when predicting financial performance of the business (Capon et al., 1990; Hendrickson and Psarouthakis, 1998). Furthermore, it is argued in the literature that ownership group size may influence performance of the business although the direction of this relationship is not clear. In the group dynamics literature, some researchers indicate the disadvantages of a larger group such as free riding (Smith and Mackie, 2000), whereas others focus on advantages of larger groups, including grater and more diversified resources available to the business through the group (Dyer, 2006; Sirmon and Hit, 2003). Without predicting either direction, we also control thus for size of the ownership group. METHOD Sample and data collection Data were collected by telephone interview in summer 2009 from a random sample of Dutch private firms registered in the Dutch Chamber of Commerce. With respect to sector, the sample was drawn proportionally according to the distribution of sectors in the overall Dutch population of firms. With respect to size, the sample is stratified such that firms were drawn equally from each of the following size classes (resulting in intentional oversampling of larger firms): 2-9; 10-49; 50-99; 100-199; 200 and more employees (including the director). Self-employed individuals (i.e. firms with one employee, in total) were thus excluded. A key informant approach was adopted (Kumar et al., 1993) by conducting a telephone interview with the managing director. Eventually a panel of 1469 companies was created, resulting in a response rate of 42%. Since some independent variables examined in this study relate to interpersonal relationships, and for which the director presumably has direct knowledge, we included only those firms for which respondents reported ownership of between two and twenty owners, leaving a subsample of 795 firms.i Further limitation of sample size to 784 cases resulted from missing values for Company Age variable. Eventually, the data for 560 companies were available for the analyses due to the missing data for some of the independent variables included in the model. A more detailed description of the sample is shown in the Appendix. Variables Shared vision. To measure shared vision amongst owners, each respondent was asked to indicate the extent to which he/she agrees or disagrees with the following statements: The owners of this business have a commitment to managing wealth as a group rather than as individuals; owners share the same vision about the business; owners are committed to growing vs. harvesting the business; and owners agree about the key objectives of the businesses. All questions were measured on a 5-point Likert scale ranging from 1=strongly disagree and 5=strongly agree. Quality of relationships. This variable was measured by asking each respondent to assess the extent to which he/she agrees or disagrees with the following statements: Owners of this business tend to trust one another; owners are open and honest with one another; owners have good cooperative relationships; and owners work together as a team. The

respondents could indicate the extent of their agreement on a 5-point Likert scale ranging from 1=strongly disagree to 5=strongly agree. Family involvement. Family involvement was measured on three indicators, including: A family relation exists between owners; there is a considerable influence on the business strategy by one family; and the business is self-described as a family business. The scale for each item is: 1=yes, 0=no. The early research in family business field considers the majority of ownership in hands of one family as an important criterion to distinguish family versus nonfamily firms (Stoy Hayward, 1989). This approach, however, is problematic as most small and medium-sized firms fulfill those criteria and can thus be defined as family firms (Uhlaner, 2005), limiting the variance to explain family effects (Brockhaus, 1994). Using multiple items can overcome this problem. Drawing from previous research (Uhlaner, 2005), in this study the above-mentioned items were used to measure the family involvement construct. Financial performance. In order to measure the dependent variable, each respondent was asked to indicate his/her perceptions of firm performance on three indicators reflecting short-term financial performance of the firm. Specifically, the indicators include 1) a firm’s financial performance compared to competitors, 2) profitability in the last fiscal year, and 3) current liquidity (Hendrickson and Psarouthakis, 1998; Sorenson, 1999; Uhlaner et al., 2007). All questions were measured on the Likert scale. For the first question the respondents could compare their financial performance with the competitors on a scale from 1=much worse to 5=much better. The second question was measured on the 7-point scale, ranging from 1=extremely unprofitable to 7=extremely profitable last fiscal year. Finally, current liquidity was assessed on a scale from 1=very little liquidity to 4=significant liquidity. Ownership-management overlap. For the moderating variable, overlap between ownership and management, the respondent was asked to indicate how many managers are also owners. The variable is constructed as a percentage of overlap, where 0% designates no overlap and 100% designates complete overlap between ownership and management. For the purposes of analysis, two groups were formed where Group 1 includes firms with partial to no overlap (i.e. overlap between 0 and 99%) and Group 2 includes firms with complete overlap (i.e. 100% overlap).ii Control variables. The control variables used in this study include: Company size, age, sector, and number of owners. Company size is measured by number of employees employed by the business both in the Netherlands and abroad in 2009. To measure company age, each respondent was asked to indicate the year when the business was originally established. On the basis of this information the business age as of 2009 was calculated. Based on BIK-codesiii, companies in the sample were grouped into five main sectors: Manufacturing, construction, wholesale & retail, agriculture, and services (including hospitality, transportation services, financial services, business services, and other services). To measure number of owners, respondents were asked to indicate how many individual owners there were. For both company size and number of owners, values were converted to natural logarithms due to skewedness in the distributions. Data analysis This study employs the Structural Equation Modeling (SEM) method of data analysis with

a maximum likelihood estimation (MLE) to analyze proposed relationships. This technique is more appropriate than regression analysis as SEM allows one to simultaneously estimate the relationships between multiple independent and dependent constructs and incorporates measurement error directly into the model (Byrne, 2010). We conduct the analysis in two steps: In the first step we assess the reliability and validity of the measurement model, and in the second step we assess the structural model (Anderson and Gerbing, 1988). To develop and test the measurement model, we first conduct an Exploratory Factor Analysis (EFA) using principal component factor analysis with Varimax rotation, to develop an initial model of dimensionality of the constructs. Next we carry out a Confirmatory Factor Analysis (CFA) using the AMOS software program, to assess the fit of the measurement model, and to check for convergent validity, construct reliability and discriminant validity (Byrne, 2010). In step two we assess the structural model to test the validity of the hypotheses. The model’s goodness-of-fit is assessed on various indicators including the chi-square (χ2 ), chisquare/degrees of freedom (χ 2/df), the comparative fix index (CFI), the goodness-of-fit index (GFI) and the root mean square error of approximation (RMSEA). We use cut-offs recommended by Byrne (2010) and Hair et al. (2006) for those indicators (χ2/df≤3.00, CFI≥.92, GFI≥.90, RMSEA≤.05). RESULTS Descriptive statistics and bivariate relationships The bivariate correlation coefficients, the mean and standard deviation for the variables are presented in Table 1. The dependent variable, financial performance, is positively correlated with company size (r=.29, p<.001), age (r=.11, p<.05), service sector (r=-.10, p<.05), number of owners (r=.12, p<.05), shared vision (r=.18, p<.01) and quality of relationships (r=.13, p<.05). The firms included in the study are on average 41 years, employ 172 employees and have 3 owners. Measurement model Exploratory factor analysis. The exploratory factor analysis presented in Table 2 indicates a four factor solution. In the rotated Varimax solution all items that are included in scales measuring quality of relationships, shared vision, family involvement and financial performance clearly load on separate factors, with all loadings equal .70 or higher, a cut-off recommended by Hair et al. (2006) as an indication of well-defined structure. The highest unintended loading equals .40, which is well below the recommended cut-off mentioned above. All four factors explain 70.92% of total variance with the first factor accounting for 34.94% of the explanation. Those findings support the preliminary discriminant validity of the measurement. Furthermore, the Cronbach’s Alpha reliability coefficients are close to or above .70, which indicates the consistency of the scales. Confirmatory factor analysis. To confirm the validity of the four-factor structure (i.e, financial performance and the three social capital variables), we compared it to two alternative models using SEM. The first alternative (Model 2) tests a three-factor solution, combining shared vision and quality of relationships into one factor. Model 3 tests a two factor solution combining all three dimensions of social capital in one factor. Table 3 presents the Chi-square difference tests and Akaike Information Criterion (AIC) for each

model, both of which clearly confirm Model 1 as the best fit for the data.iv Moreover, based on the various goodness-of-fit indicators we may conclude that this model fits the observed data well (χ2=166.760, df=71, p<.001, χ2/df=2.349, GFI=.961, CFI=.975, RMSEA=.049).v Convergent validity. To assess convergent validity, we assess the factor loadings from the CFA, the average variance extracted (AVE) by each factor, and construct reliability (CR) (Hair et al., 2006). All standardized loadings meet or exceed the .50 cut off for practical significance recommended by Hair et al. (2006), with the majority exceeding the more stringent .70 cut-off. Moreover, all loadings are significant at the p<.001 level. The composite reliabilities for quality of relationships, shared vision and family involvement range from .80 to .94, all of which exceed the recommended cut off of .70 (Hair et al., 2006). The composite reliability for financial performance is slightly below this suggested value (CR=.69) but still acceptable (Hair et al., 2006). Finally, the AVE for the constructs measuring ownership social capital are above the .50 cut off (Hair et al., 2006), whereas this index for financial performance is somewhat lower (AVE=.43). Taking those results together we can conclude that, despite weaker indicators for financial performance, results indicate reasonable convergent validity for each of the four constructs. Discriminant validity. To assess the discriminant validity, we compared the square root of AVE for each construct (see Table 1), with the correlation between this construct and other constructs (Hair et al., 2006). Results support discriminant validity for the four proposed constructs, given that the square root of AVE is higher than the inter-construct correlations. Structural model Direct effect of shared vision, quality of relationships and family involvement. The measurement model developed in step one of the data analysis is used to investigate the hypothesized influence of shared vision, quality of relationships and family involvement on financial performance of the firm.vi Figure 2 represents the results for the whole sample. The model fits with the observed data (χ2=387.193, df=165, p<.001, χ2/df=2.347, GFI=.938, CFI=.949, RMSEA=.049) and explains 15% of variance in the dependent variable (9.5% of which is due to control variables, 5.5% added variance of which is due to the ownership social capital variables).vii However, in the baseline model, the paths between none of the three independent variables and the dependent variable, financial performance, is significant (see Figure 2). Amongst the control variables only company size predicts the dependent variable (β=.28, p<.001). Taken together, these results do not support Hypothesis 1, 2 or 3.

TABLE 1 Assessment of Discriminant Validity as well as Correlations, Means and Standard Deviations for the Variables used in the Analysis 2 3 4 5 6 7 8 9 10 11 12 1 1. Financial Performance 1 2. Company Size (ln) .29c 1 a 3. Company Age .11 .29c 1 4. Service -.10a -.12b -.26c 1 5. Agriculture -.05a -.07x .00x -.13b 1 x c c c 6. Manufacturing .04 .22 .23 -.42 -.07x 1 x x a c 7. Construction .05 .00 .11 -.30 -.05x -.18c 1 x x x c a c 8. Wholesale & Retail .06 -.05 -.00 -.48 -.09 -.29 -.21c 1 9. No. of owners (ln) .12a .36c .06x .09a -.01x .04x -.08x -.07x 1 x a c c b x x c 10. Family Involvement .07 -.09 .21 -.19 .15 -.02 -.01 .19 -.18c 1 11. Shared Vision .18b -.01x .07x -.07x .03x .03x .01x .04x -.12a .02x 1 a b x x x x x x c b 12. Quality of Relationships .13 -.12 .02 -.02 .06 -.05 -.00 .04 -.21 .14 .71c 1 MEAN d, e -.02 172.41 40.81 .42 .02 .20 .11 .25 3.34 .54 4.23 4.41 SD 1.00 662.24 36.15 .49 .15 .40 .32 .43 2.63 .42 .53 .60 Square root of AVE .66 - .76 .72 .89 a b c : p<.05; : p<.01; : p<.001; N=560 d : The mean and standard deviation for Financial Performance is calculated for the object score for the overall sample. e : The mean and standard deviation for all variables except Financial Performance are reported on the base of unconverted values.

TABLE 2 Exploratory Factor Analysisa Label Item Quality of Shared Relationships Vision QR1 Owners tend to trust one another .26 .87 QR2 Owners are open and honest with one another .27 .88 QR3 Owners have good cooperative relationships .27 .88 QR4 Owners work together as a team .26 .85 SV1 Owners are committed to managing wealth as .12 .80 a group rather than as individuals SV2 Owners share the same vision about the business .40 .71 SV3 Owners are committed to growing vs. harvesting .26 .70 the business SV4 Owners agree about the key objectives of .40 .73 the business FamInf1 Self-description as a family business .07 .03 FamInf2 Family relation between owners .07 -.06 FamInf3 The family has a considerable influence .03 .04 on the business strategy FinPerf1 Financial performance as compared to major .01 .00 competitors in industry FinPerf2 The profitability in the last fiscal year .05 .09 FinPerf3 Current liquidity .05 .03 Percentage variance explained 34.94 15.32 Cronbach-Alpha reliability coefficientsb .93 .80 a : Principal Component Analysis with Varimax Rotation, 4 components extracted; N=560; b : Unstandardized reliability coefficients are reported. NOTE: Bold values indicate items with the highest loadings for the specific factor.

Family Involvement .06 .06 .02 .11 .06

Financial Performance .06 .04 .01 .05 -.01

.01 -.03

.03 .08

-.04

.09

.84 .88 .80

.09 -.03 .01

-.02

.78

.03 .06 12.86 .80

.81 .75 7.80 .66

TABLE 3 Testing the Fit for Alternative Measurement Models using Confirmatory Factor Analysis df ∆χ2 (∆df) AIC χ2 index Model 1: 4-factors: shared vision, 166.76 71 234.76 quality of relationships, family involvement plus financial performance Model 2: 3-factors: shared vision 486.60 74 319.82 (3)c 548.60 and quality of relationships combined; family involvement plus financial performance Model 3: 2-factors: all three 1015.90 76 849.14 (5)c 1073.90 ownership social capital dimensions combined into one dimension plus financial performance c : p<.001. TABLE 4 Confirmatory Factor Analysis Construct Measurement Standardized Composite item loading Reliability (CR) Quality of Relationship QR1 .89 .94 QR2 .91 QR3 .90 QR4 .84 Shared Vision SV1 .60 .81 SV2 .79 SV3 .63 SV4 .83 Family Involvement FamInv1 .74 .80 FamInv2 .87 FamInv3 .66 Financial Performance FinPerf1 .61 .69 FinPerf2 .76 FinPerf3 .58 NOTE: Standardized factor loadings significant at p<.001; N=560.

Average Variance Extracted (AVE) .78

.52

.58

.43

Ownership-management overlap as a moderator variable. Hypothesis 4 proposes that ownership-management overlap moderates the relationship between each of the independent variables and the dependent variable. To test this hypothesis we first carry out analyses for the baseline SEM model (see Figure 2) for the two groups defined in the method section. The indicators of model fit of the two-group model allow us to conclude that the model fits the observed data (χ2 =580.294, df=330, p<.001, χ2/df=1.758, GFI=.912, CFI=.943, RMSEA=.037). Figure 3 presents results for Group 1 (partial to no overlap). Only family involvement predicts financial performance (β=.22, p<.01). Shared vision and quality of relationships are not significantly related to financial performance (β=-.04, ns and β=.17, ns, respectively). The model predicts 19% of the dependent variable (see R2 reported in Figure 3), whereas controls predict only 8.5% of financial performance (model not shown in figures), reflecting ∆R2 of .105 for the three ownership social capital variables. Figure 4 presents the structural model for Group 2 (complete overlap). In contrast to the structural model for Group 1, the path from shared vision to financial performance is positive and significant (β=.35, p<.01), whereas neither quality of relationships nor family involvement are related to dependent variable (β=-.11, ns and β=-.08, ns, respectively). This model explains 23% the variance in the dependent variable compared with 8.5% for a model of control variables only, reflecting ∆R2 of .145 for the three ownership social capital variables. In order to test the significance of the differences in the structural models shown in Figures 3 and 4, for Group 1 and Group 2, we conducted the critical ratios of differences test. The test revealed that the differences in relationships between two of the three independent variables and the dependent variable are significantly different, with a trend for the third variable (for shared vision, z= 2.259, p<.05; for quality of relationships, z= 2.819, p<.01; for family involvement, z=-1.758, p<.10). Further test of moderation effect. Since the hypothesized relationships for Group 1 and Group 2 differ significantly but a number of paths in each model is not significant (see Figure 3 and Figure 4), we further explore post hoc models for each group with good fit but fewer nonsignificant paths (see Byrne, 2010). Figure 5 shows an improved post hoc model for Group 1. Although this model explains the same amount of the variance in financial performance as the model presented in Figure 2 (R2 =.19 in both models), by removing nonsignificant path between shared vision and financial performance we obtained significant results for paths between both quality of relationships and financial performance (β=.14, p<.05) as well as family involvement and financial performance (β=.22, p<.01). However, shared vision predicts quality of relationships (β=.69, p<.001), suggesting an indirect effect on the dependent variable. Finally, a positive and significant path from family involvement to quality of relationships (β=.12, p<.05) suggests that in addition to a direct effect, family involvement also indirectly affects financial performance by way of quality of relationships. The model has a slightly worse fit then the model in Figure 3, but still the goodness-of-fit indexes meet the requirements for a well-fitted model (χ 2=616.496, df=336, p<.001, χ2/df=1.835, GFI=.908,

CFI=.936, RMSEA=.039). Additionally, we conducted the critical ratios of differences test for this model between the firms in Group 1 and Group 2. We find a significant difference only for the relationship between family involvement and financial performance (z=-3.309, p<.01). FIGURE 2 Baseline Structural Model (Full Sample)

Shared Vision .13, ns

.71c

R2=.15

.00, ns .07, ns

Quality of Relationships

Financial Performance of the Business

.09, ns .12

b

Family Involvement

a

: p<.05; b: p<.01; c: p<.001; N=560 NOTES: Standardized regression weights reported. Correlations between exogenous variables italicized. Control variables: company age, size, sector and number of owners. Χ2=387.193, df=165, p<.001, χ2/df=2.347, GFI=.938, CFI=.949, RMSEA=.049.

FIGURE 3 Baseline Model: Group 1: Firms with Partial to Overlap between Ownership and Management

Shared Vision

.69

-.04, ns

c

R2=.19

-01, ns .17, ns

Quality of Relationships

Financial Performance of the Business

.22b .12, ns

Family Involvement a

: p<.05; b: p<.01; c: p<.001; N=301. NOTE: Standardized regression weights reported. Correlations between exogenous variables italicized. Control variables: company age, size, sector and number of owners. Χ2=580.294, df=330, p<.001, χ2/df=1.758, GFI=.912, CFI=.943, RMSEA=.037. FIGURE 4 Baseline model: Group 2: Firms with Complete Overlap between Ownership and Management

Shared Vision .35b

.71c

R2=.23

-.02, ns -.11, ns

Quality of Relationships

-.08, ns .11, ns

Family Involvement

Financial Performance of the Business

a

: p<.05; b: p<.01; c: p<.001; N=259. NOTE: Standardized regression weights reported. Correlations between exogenous variables italicized. Control variables: company age, size, sector and number of owners. Χ2=580.294, df=330, p<.001, χ2/df=1.758, GFI=.912, CFI=.943, RMSEA=.037.

We also developed an improved post hoc model for Group 2. From the standpoint of eliminating nonsignificant paths while retaining goodness-of-fit, the best model we could obtain is presented in Figure 6. Note that in this model, the dependent variable is directly predicted by shared vision (β=.27, p<.001). Furthermore, the model shows a double mediation such that family involvement predicts financial performance via quality of relationships and shared vision, and shared version mediates the relationship between quality of relationships and financial performance.viii The model fits the observed data well (χ2=627.316, df=338, p<.001, χ2/df=1.856, GFI=.906, CFI=.934, RMSEA=.039). Interestingly, the critical ratios of differences test for this model shows, that there are no differences between Groups 1 and 2. Nevertheless, the paths for Group 1 are nonsignificant, except for the path between quality of relationships and shared vision (β=.70, p<.001). FIGURE 5 Improved Post hoc Model for Group 1: Firms with Partial to no Overlap between Ownership and Management

Shared Vision

R2=.19

.69c Quality of Relationships

.14a

Financial Performance of the Business

.22b .12a

Family Involvement

a

: p<.05; b: p<.01; c: p<.001; N=301. NOTE: Standardized regression weights reported. Control variables: company age, size, sector and number of owners. Χ2=616.496, df=336, p<.001, χ2/df=1.835, GFI=.908, CFI=.936, RMSEA=.039.

FIGURE 6: Improved Post hoc Model for Group 2: Firms with Complete Overlap between Ownership and Management Shared Vision

.71

.27c

c

R2=.21 Financial Performance of the Business

Quality of Relationships .14a Family Involvement

a

: p<.05; b: p<.01; c: p<.001; N=259. NOTE: Standardized regression weights reported. Control variables: company age, size, sector and number of owners. Χ2=627.316, df=338, p<.001, χ2/df=1.856, GFI=.906, CFI=.934, RMSEA=.039. DISCUSSION AND CONCLUSIONS Initial discussion The aim of this paper was to explore the influence of ownership social capital on business performance in the context of privately firms owned by small to medium sized groups of owners. Three aspects of ownership social capital are studied, including the relational dimension of ownership social capital, represented by the quality of relationships between owners, the cognitive dimension of ownership social capital, represented by shared vision amongst owners, and structural dimension of ownership social capital, represented by family involvement. Contrary to results from a recent study by Carr et al. (2011) for family social capital, we clearly find support for a three factor solution for ownership social capital. Closer examination of items selected for Carr et al.’s study, however, would suggest that their interpretation of structural social capital is quite different than in our own study. ix Regarding our hypotheses, the baseline model provides weak support for the first three hypotheses. None of the three variables, individually predicts financial performance, although together, they explain an additional 5.5% variance after the control variables. Furthermore, when the baseline model is used initially in the test for moderator effects, only Hypothesis 4c (linking family involvement and financial performance for the partial to no overlap group), is supported. Counter to Hypotheses 4a and 4b, neither shared vision nor quality of relationships directly predicts financial performance in the same baseline

model. However, when post hoc analyses are carried out for subgroups of firms, much stronger support for direct and indirect effects of ownership social capital emerge, although patterns are quite different for the two groups. In Group 1, there are one or more managers who are not owners. That under such conditions, family involvement has a more positive effect on financial performance can be explained by the agency theory and family social capital arguments described in the introduction. On the one hand, managers with no family ties to the owners may be expected to act more selfishly than those managers who either feel a loyalty of kinship to the owners or even further, may expect one day to inherit the business. On the other hand, in conditions of complete overlap between managers and owners, all managers will have direct benefits from serving the entire firm since they are of course, also owners. Thus, in the second group, there is no difference in the motives for the family vs. nonfamily managers. With respect to shared vision, it may be that it has a more direct effect on financial performance in the second group, since the owner-management group has more direct influence on the implementation of such objectives. Theoretical and practical implications of the findings This study has several theoretical and practical implications. From a theoretical perspective, we broaden the application of the literature on social capital by applying this concept in the context of the ownership group. Moreover, we enhance the discussion about the possible effect of social capital on organizational effectiveness by empirically supporting the notion that under certain conditions, cognitive, relational, and structural social capital, as represented by shared vision, quality of relationships and family involvement, enhance financial performance of the firm, either directly or indirectly, depending on the degree of overlap between ownership and management. Furthermore, although the relationship between different dimensions of organizational social capital is not clear in the conceptual studies (Adler and Kwon, 2002; Leana and Van Buren, 1999; Nahapiet and Ghoshal, 1998), we find that the relational dimension of ownership social capital serves as a mediator between cognitive dimension of ownership social capital and business performance but most markedly for firms with partial to no overlap between ownership and management. The pattern seems reversed for firms with complete overlap. That is, shared vision serves under these conditions as the mediating variable. Furthermore, our study also enhances the discussion in the family business field by demonstrating that family involvement can have a direct positive influence on financial performance of the business, but especially under conditions where at least some managers are not owners. Past research has not used this variable as a moderator (e.g. O’Boyle et al., 2012). However, the significant differences in structural models in our research clearly indicate the need to do so before drawing conclusions about the effects of family involvement on financial performance. This study also has implications for practitioners. This empirical research supports the argument that owners – as one of the social entities present in the business – can indeed influence financial performance of the firm. Specifically, good relationships between owners seem to be essential since they directly influence firm performance and other variables. A shared vision is also beneficial to the business, but only indirectly improving financial performance via the quality of relationships amongst owners when at least some owners are not managers. Regardless, findings suggest that under the right conditions, an

effective owning group can be a valuable resource for the firm. Limitations of the research and recommendations for future research The current research is not free from limitations. Firstly, we could improve on the scale measuring financial performance. Although we use techniques most appropriate for cross sectional data, with respect to proposed paths, we cannot rule out that the causal direction between financial performance and other variables may be reversed. Collecting data for a later time period for the same panel could strengthen our conclusions in future research. Another limitation of our study is that in spite of the statistically significant findings, in absolute terms, a rather small incremental percentage of variance in the dependent variable is explained by the proposed models. Future research might explore additional ownership dynamics variables, which can help to explain a larger portion of variation in financial performance or other outcomes. However, examining subsamples based on the degree of overlap between ownership and management greatly enhanced the predictive power of our models (adding between 5% and 9% variance in the dependent variable explained by ownership social capital variables over and above the amount (5.5%) explained by the three variables in the baseline model, depending on the subgroup in question). These results suggest that ownership-management overlap is an extremely important variable to consider in corporate governance research, not only as a control, but also as a moderator variable. As discussed in the paper, social capital had been conceptualized as having three dimensions: Relational, cognitive and structural (Nahapiet and Ghoshal, 1998). The present study, however, does not include the measurement of structural social capital reflected in the types and characteristics of linkages that owners have with external parties. We recommend that future research examine the external linkages of ownership group and their consequences in more detail. In particular, we draw attention to the question of how “weak ties” (i.e. casual contacts) and “strong” ties (i.e. close family and friends) (Granovetter, 1982) available to the owners are used by them, and what influence such ties have for business success. Future research could also investigate the effect of various human capital variables of the owners (e.g. education, work experience, diversity), to see how this may alter the effects of the social capital variables. Finally, future research would benefit from collecting data from multiple observers per firm, which would improve reliability and validity of the measures. Conclusions The aim of this paper was to explore whether ownership social capital i.e., shared vision amongst owners, the quality of relationship between owners, and family involvement influences business performance. The results of this empirical study conducted on 560 private firms with between two and twenty owners, show that owners are indeed important actors since the relationships between them influence business success. Thus, we would welcome more research on the ownership group when exploring factors influencing organizational success. This study shows the importance of building unified groups of owners who work as a team towards a common goal. Although the types of effects (direct versus indirect) vary according to the degree of overlap between ownership and management, our results suggest

that regardless of overlap, both the quality of relationships and shared vision amongst owners can have positive effects on financial performance. REFERENCES Adler, P.S. and Kwon, S.W. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27 (1), 17-40. Anderson, J.C., and Gerbing, D.W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411-423. Anderson, R., and Reeb, D. (2003). Founding family ownership and firm performance: Evidence from the S&P 500. Journal of Finance, 58 (3), 1301-1328. Arregle, L.J., Hitt, M.A., Sirmon, D.G., and Very, P. (2007). The development of organizational social capital: Attributes of family firms. Journal of Management Studies, 44 (1), 73-95. Berent-Braun, M.M., and Uhlaner, L.M. (2012). Family governance practices and teambuilding: Paradox of the enterprising family. Small Business Economics, 38(1), 103-119. Bolino, M.C., Turnley, W H., and Bloodgood, J.M. (2002). Citizenship behavior and the creation of social capital in organizations. Academy of Management Review, 27 (4), 505-522. Bosma, N., Van Praag, M., Thurik, R. and De Wit. G. (2004). The value of human and social capital investments for the business performance of startups. Small Business Economics, 23, 227-236. Bourdieu, P. (1986). The forms of capital. In: J.G. Richardson (Ed.), Handbook of the theory and research for the sociology of education, pp. 241-258. New York: Greenwood. Bouty, I. (2000). Interpersonal and interactional influences on informal resources exchanges between R&D researchers across organizational borders. Academy of Management Journal, 43 (1), 50-65. Brockhaus, R.H. (1994). Entrepreneurship and family business research: Comparisons, critique and lessons. Entrepreneurship: Theory & Practice, 19 (1), 25-38. Browne, M.W., and Cudeck, R. (1989). Single sample cross-validation indices for covariance structures. Multivariate Behavioral Research, 24(4), 445-455. Byrne, B. (2010). Structural equation modeling with AMOS: Basic concepts, applications and programming, 2nd edition. New York: Routledge, Taylor and Francis Group. Capon, N., Farley, J. U., and Hoenig, S. (1990). Determinants of financial performance: A metaanalysis. Management Science, 36, 1143-1155. Carr, J.C., Cole, M.S., Ring, J.K., and Blettner, D.P. (2011). A measure of variations in internal social capital among family firms. Entrepreneurship: Theory & Practice, 35 (6), 1207-27. Cohen, S.G., and Bailey, D.E. (1997). What makes teams work: Group effectiveness research from the shop floor to the executive suite. Journal of Management, 23, 239-90. Coleman, J.S. (1990). Foundations of social theory. Cambridge, MA: Belknap Press of Harvard University Press. Combs, J., Liu, Y., Hall, A., and Ketchen, D. (2006). How much do high-performance practices matter? A meta-analysis of their effects on organizational performance. Personnel Psychology, 59, 201-528. Dess, G.G. (1987). Consensus on strategy formulation and organizational performance: Competitors in a fragmented industry. Strategic Management Journal, 8 (3), 259-277. Dyer, W.G. (2006). Examining the ‘family effect’ on firm performance. Family Business Review, 19, 253-273.

Ensley, M.D. and Pearson, A.W. (2005). An exploratory comparison of the behavioral dynamics of top management teams in family and nonfamily new ventures: Cohesion, conflict, potency, and consensus. Entrepreneurship: Theory & Practice, 29 (3), 267-284. Etzioni, A. (1996). The responsive community: A communitarian perspective. American Sociological Review, 61 (1), 1-11. Granovetter, M.S. (1982). The strength of weak ties: A network theory revised. In: P.V. Marsden and N. Lin (Eds.), Social structure and network analysis, pp. 105-130. Beverly Hills, CA: Sage. Guth, W.D., and MacMillan, I.C. (1986). Strategic implementation versus middle management self-interest. Strategic Management Journal, 7 (4), 313-327. Guthrie, J. P. (2001). High-involvement work practices, turnover, and productivity: Evidence from New Zealand. Academy of Management Journal, 44, 80-190. Habbershon, T.G., and Williams, M.L. (1999). A resource-based framework for assessing the strategic advantages of family firms. Family Business Review, 12 (1), 1-25. Habbershon, T.G., Williams, M.L., and MacMillan, I.C. (2003). Familiness: A unified systems theory of family business performance. Journal of Business Venturing, 18, 451-65. Hair, F.J., Black, W.C., Babin, B.J., Anderson, R.E., and Tatham, R.L. (2006). Multivariate Data Analysis. New Jersey, Upper Saddle River: Prentice Hall. Hendrickson, L.U., and Psarouthakis, J. (1998). Dynamic management of growing firms: A strategic approach (2nd ed.). Ann Arbor: The University of Michigan Press. Hitt, M.A., Lee, H., and Yucel, E. (2002). The importance of social capital to the management of multinational enterprises: Relational networks amongst Asian and western firms. Asia Pacific Journal of Management, 19 (2/3), 353-372. Huszczo, G.E. (1996). Tools for team excellence: Getting your team into high gear and keeping it there. Palo Alto, CA: Davies-Black Publishing. Jensen, M.C., and Meckling, W.H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305-360. Jones, G.R., and George, J.M. (1998). The experience and evolution of trust: Implications for cooperation and teamwork. Academy of Management Review, 23 (3), 531-546. Kumar, N.L. , Stern, W., and Anderson, J.W. (1993) Conducting interorganizational research using key informants. Academy of Management Journal, 36, 1663-1651. Lazega, E., and Pattison, P.E. (2001). Social capital as social mechanisms and collective assets: The example of status auctions among colleagues. In: N. Lin, K.C. Cook, and R.S. Burt (Eds.), Social capital: Theory and research. New York: Aldine de Gruyter, 185-208. Leana, C.R and Van Buren, H.J. (1999). Organizational social capital and employment practices. The Academy of Management Review, 24 (3), 538-555. Lee, R. (2009). Social capital and business and management: Setting a research agenda. International Journal of Management Reviews, 11 (3), 247-273. Lin, N. (2001). Social capital: A theory of social structure and action. New York: Cambridge University Press. Loury, G.C. (1977). A dynamic theory of racial income differences. In P.A. Wallace & A. M. LaMonde (eds.), Women, minorities and employment discrimination. Lexington, NA: Lexington Books, 153-186. Madsen, E.S., Smith, V., and Dilling-Hansen, M. (2007). Why does firm have several owners? Corporate Ownership & Control, 4 (3), 80-86.

McEvily, B., Perrone, V., and Zheer, A. (2003). Introduction to the special issue on trust in an organizational context. Organization Science, 14 (1), 1-4. Moreno, A.M., and Casillas, J.C. (2008). Entrepreneurial orientation and growth of SMEs: A casual model. Entrepreneurship: Theory & Practice, 32 (3), 507-528. Mustakallio, M.E., Autio, E., and Zahra, S.A. (2002). Relational and contractual governance in family business: Effects on strategic decision making. Family Business Review, 15 (3), 205222. Nahapiet, J. and Ghoshal, S. (1998). Social capital, intellectual capital and the organizational advantage. Academy of Management Review, 23 (2), 242-266. O’Boyle, E.H., Pollack, J.M., and Rutherford, M.W. (2012). Exploring the relation between family involvement and firm’s financial performance: A meta-analysis of main moderator effects. Journal of Business Venturing, 27, 1-18. Pedersen, T., and Thomsen, S. (2003). Ownership structure and value of the largest European firms: The importance of owner identity. Journal of Governance and Management, 7, 27-55. Putnam, R. (1993a). The prosperous community: Social capital and public life. American Prospect, 13 (1), 35-42. Putnam, R. (1993b). Making democracy work: Civic traditions in modern Italy. Princeton, NJ: Princeton University Press. Sirmon, D.G., and Hitt, M.A. (2003). Managing resources: Linking unique resources, management, and wealth creation in family firms. Entrepreneurship: Theory & Practice, 27, 339-358. Smith, E.R., and Mackie, D.M. (2000). Social psychology (2nd ed.). Philadelphia, PA: Psychology Press. Sorenson, R.L. (1999). Conflict management strategies used in successful family businesses. Family Business Review, 12 (2), 133–146. Stoy Hayward. (1989). Staying the course: Survival characteristics of the family owned business. Stoy Hayward: London. Thomsen, S., and Pedersen, T. (2000). Ownership structure and economic performance in the largest European companies. Strategic Management Journal, 21, 689-705. Uhlaner, L.M. (2005). The use of the Guttman scale in development of a family orientation index for small-to-medium-sized firms. Family Business Review, 18, 41-56. Uhlaner, L.M., Flören, R.H., and Geerlings, J.R. (2007). Owner commitment and relational governance in the privately-held firm: An empirical study. Small Business Economics, 29 (3), 275-293. Walker, G., Kogut, B., and Shan, W. (1997). Social capital, structural holes and the formation of an industry network. Organization Science, 8 (2), 109-125. Weick, K.E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage. Westlund, H., and Adam, F. (2010). Social capital and economic performance: A meta-analysis of 65 studies. European Planning Studies, 18(6), 893-919. Williams, R.O. (1992). Successful ownership in business families. Family Business Review, 5 (2), 161-172. Wu, W. (2008). Dimensions of social capital and firm competitiveness improvement: The mediating role of information sharing. Journal of Management Studies, 45 (1), 122-146. Wu, Z., Hedges, P.L., and Zhang, S. (2007). Effects of concentrated ownership and owner management on small business debt financing. Journal of Small Business Management, 45 (4), 422-437.

APPENDIX More detailed description of the Sample Full sample Final estimation sample Selection sample * ** (N=784) (N=560)*** (N=619) Number Percentage Number Percentage Number Percentage of firms of firms of firms of firms of firms of firms Company size 2–9 10 – 49 50 – 99 100 – 199 200 and more Total

187 187 159 157 94 784

23.9 23.9 20.3 20.0 12.0 100.0

131 142 125 136 85 619

21.2 22.9 20.2 22.0 13.7 100.0

116 131 111 121 81 560

20.7 23.4 19.8 21.6 14.5 100.0

Company Age 1 – 4 years 5–9 10 – 19 20 – 49 50 – 99 100 and older Total

67 89 140 257 162 69 784

8.6 11.4 17.9 32.8 20.7 8.8 100.0

48 61 107 208 139 56 619

7.8 9.9 17.3 33.6 22.5 9.0 100.0

45 55 99 185 126 50 560

8.0 9.8 17.7 33.0 22.5 8.9 100.0

Sector Service Agriculture Manufacturing Construction Wholesale & Retail Total

325 22 146 100 192 784

41.4 2.8 18.6 12.8 24.4 100.0

256 16 121 71 155 619

41.4 2.6 19.5 11.5 25.0 100.0

234 12 112 64 138 560

41.9 2.1 20.0 11.4 24.6 100.0

Number of owners 2 444 56.6 335 54.1 302 53.9 3 151 19.3 121 19.5 111 19.8 4–9 166 21.2 141 22.8 125 22.3 10 – 20 23 2.9 22 3.6 22 3.9 Total 784 100.0 619 100.0 560 100.0 *** This sample is composed of all cases with 2 to 20 private owners with all data on control variables. *** This sample is composed of a subgroup of firms from the full sample for which the financial performance data is available. ***

This sample is used for the estimation of proposed models. Further limitation of the sample is due to the missing data for one or more independent variables.

ENDNOTES i

The following cases are excluded in this sample: cases with 1 or more than 20 owners (N=557), cases with an institution as an owner (N=65), cases for which the exact number of owners is not known (N=50), and cases with missing data for Number of Owners (N=2). ii Since we found no significant differences between firms with 0% overlap and firms with overlap larger than 0% but smaller than 100%, we combined those two groups into one cohort. iii The BIK-code is used by the Dutch Chamber of Commerce to classify businesses in separate industry sectors according to their main activities. iv Refer to Byrne (2010) for the technique used to select the best fitting measurement model. v Although the chi-square test should be not significant (that is, there should be no significant difference between observed data and the prediction of the model), in case of the larger samples (N>250) with more than 12 observed variables significant p-values can be expected (Hair et al., 2006). Since the analysis is conducted on the sample of 560 cases and includes 14 observed variables, we can expect that the chi-square difference test will be significant. vi In creating the structural model, we followed the technique recommended by Byrne (2010) to review modification indices and make certain adjustments to the final model. We did this for control variables only, letting SEM estimate the correlations between company size and age, company size and number of owners, as well as each of the sector dummies with each other and with company size and age. The modification indices further indicated that correlations between age and family involvement as well as age and number of owners should be estimated. Thus, these were also added. vii This was calculated by comparing variance explained for a SEM structural model with control variables only and a second model including all variables in the model. viii Family involvement does not predict shared vision (β=-.07, ns). ix For instance, their structural items reflect the quality of communication amongst family members.