Family control and influence on JV investment – the moderating effect of JV type and IC components

Purpose – Thispaperaimstoinvestigatetheeffectofthenatureofownershipandboardcharacteristicsonthe investment choices in joint ventures (JVs) from the dimensional point of view, controlling for the effect of JV type and other components of intellectual capital. Design/methodology/approach – The authors study a sample of Italian, Spanish, German and French nonfinancial listed firms over the 2010 – 2018 period, controlling for the fixed effects of the company ’ s sector of operation and the year. The authors also analyze the effect of family control and influence on JV investment size, taking into consideration certainboard characteristics, the typeof JV, human capital efficiency, structural capital efficiency and capital employed efficiency while also controlling for a firm ’ s profitability and size. To test the hypotheses, GLS panel data was used. Findings – The results indicate that the size of the investment in JVs is smaller for family firms than for nonfamilybusinesses.ThepresenceofCEOdualityhasanopposingeffectonthesizeoftheinvestmentinjointventuresasithasaloweringeffectinfamilybusinesseswhileitexertsanamplifierinfluenceinnonfamilybusinesses.Moreover,thetypeofjointventurehasasignificanteffectforfamilyfirms:thechoiceofalinkjointventurereducesthesizeoftheinvestment.TheauthorsfindthathumancapitalefficiencyincreasesJVinvestmentsizeforallfirms. Originality/value – This study is the first to analyze the effect of the main dimension of socioemotional wealth – family control and influence – on a firm ’ s JV investment size. It controls for the effect of JV type – link or scale – and the interplay of the other IC components.


Introduction
Family companies are a leading organizational form in economies across the world. According to the Family Firms' Institute, two-thirds of all companies around the world are family-managed and more than 70% of global GDP annually is produced by family businesses that provide 50-80% of the world's employment [1].
Engaging in new ventures is a priority for family companies trying to survive, achieve profitability and growth; a connection with other companies would allow them to gain the knowledge necessary to stimulate entrepreneurship and avoid the effect of conservatism that can threaten family firms' financial sustainability (Zahra, 2010;Basly, 2007). Family firms are characterized by a unique social capital (Arregle et al., 2007), relevant in leveraging the firm's absorptive capacity to transform and use external knowledge (Anders en, 2015). Moreover, there is evidence that alliances and partnerships, such as joint ventures (JVs), play a relevant role in forming company legitimacy and their sustainable technologies (Kishna et al., 2017;Muñoz de Prat et al., 2020).
JV literature has addressed several topics such as JV performance, innovation and internationalization. A stream of this literature has taken into account the effect of the nature of a firm's ownership on JV investment choices, mostly focusing on the internationalization perspective and entry mode choices (Sestu and Majocchi, 2020;Debellis et al., 2021;Scholes et al., 2016;Kuo et al., 2012;Claver et al., 2007), and innovation (Feranita et al., 2017). These exploratory studies have analyzed the behavior of family businesses by comparing them with nonfamily businesses as homogeneous groups. Family firms are not a homogeneous group, and their heterogeneity depends not only on companies' ownership structure but also on the degree and kind of influence the family exerts on a company's strategic decisions (Chrisman et al., 2015;Calabr o et al., 2019). The effect of ownership structure heterogeneity on JV has been analyzed by literature (Pongelli et al., 2016), but, to the best of our knowledge, literature has not yet investigated family firms' investment in JVs through the lens of the second type of heterogeneity, a relevant gap that deserves to be addressed.
Moreover, literature has pointed out the influence of the unique bundle of resources generated by the interaction of family and business (Habbershon and Williams, 1999) on a firm's intellectual capital (Cabrera-Su arez et al., 2001). As a matter of fact, family firms rely on specific human capital intangibles provided by the family's involvement in the firm and related to the founder's personality, the values shared by family members or the knowledge acquired from previous generations (Sirmon and Hitt, 2003;Claver-Cort es et al., 2015). The family's moderating role on the relationship between IC and innovation (Di eguez-Soto et al., 2016), its effect on relational capability development (McGrath and O'Toole, 2018), as well as family firms' unique characteristics in creating and governing joint ventures (Debellis et al., 2021) have been highlighted by the literature.
JVs are the expression of a peculiar component of a company's intellectual capital (IC), i.e. its relational capital that represents the part of knowledge embedded in relationships with external stakeholders (Sharabati et al., 2010;Zambon, 2002). According to the type of alliancelink or scalepartners contribute different, or similar, capabilities (Hennart, 1988) facing different risks and producing different outcomes (Dussauge et al., 2000(Dussauge et al., , 2004Khamseh and Jolly, 2014).
Recent research has also highlighted the importance of pointing out the possible interactions and synergies between the different components of IC (Crupi et al., 2020): they influence each other and certain components of IC can benefit from the improvement of other components (Johnson, 1999;Reed et al., 2006). However, to the best of our knowledge, the literature has not yet analyzed the effect of IC and JV type in family business investment in joint ventures. This paper addresses the following research question: Does family control and type of influence affect investment size in JVs? If so, does JV type and the interplay of other IC components moderate this effect?
We investigate how different types of family influence (CEO, CEO duality, number of family members on the board) affect firms' investment size in JVs, taking into account the effect of IC and JV type. The extent to which family members occupy board positions is a major source of heterogeneity in family businesses (Nordqvist et al., 2014;Di eguez-Soto and Mart ınez-Romero, 2016) and the characteristics and background of board members affect a firm's strategic choices (Hambrick, 2007;Mueller and Barker, 1997;Sciascia et al., 2013). We address the research question by basing it on the socioemotional wealth (SEW) concept. This refers to the stock of affect-related value that a family derives from its controlling position in a firm (Gomez-Mejia et al., 2007;Berrone et al., 2012). It shapes firms' decisions (Gomez-Mejia et al., 2011) and may affect investment in JVs in different ways.
We analyze a sample of Italian, Spanish, German and French nonfinancial listed firms over the period 2010-2018. The sample is of particular interest as these economies are widely Family control and influence on JV investment characterized by the presence of family businesses that have long-lasting family control (Franks et al., 2012).
Our results point out that family control and influence negatively affect JV investment size; the presence of CEO duality has an opposing effect on the size of the investment in joint ventures as it has a decreasing effect in family businesses while it exerts an amplifier influence in nonfamily businesses. The type of joint venture has a significant effect only for family businesses, whilst human capital efficiency is significant only at the level of all firms.
This study contributes to the existing research on joint ventures by analyzing the effect of JV type on family and nonfamily firms' joint-controlled investments and it adds to intellectual capital literature by also investigating the efficiency by which IC components affect the size of the investment.
Our research contributes to family-firm literature, in particular to the stream of research on the determinants of family-firm investments pointing out the effect of family influence heterogeneity, in terms of board characteristics, on JV investments.
The paper is organized as follows: Section 2 provides the theoretical framework, the literature review and the consequent hypotheses; Section 3 describes the data and methodology; Section 4 presents the results; Section 5 provides the discussion and Section 6 concludes by pointing out the contributions, implications and the limitations of the study as well as including suggestions for future research.

Theoretical framework, literature review and hypotheses 2.1 Socioemotional wealth and its dimensions
We address our research question through the lens of the socioemotional wealth (SEW) concept. Based on the behavioral agency model, developed in order to explain executives' risk-taking behavior (Wiseman and Gomez-Mejia, 1998). Gomez-Mejia et al. (2007) argue that family businesses' risk-taking behavior is determined by the effect of their decisions on the SEW and that this is why they are not always risk-averse. Further, it is also why their risk preferences are not constant (Chrisman and Patel, 2012;Kraiczy et al., 2015). The SEW refers to the stock of affect-related value that a family derives from its controlling position in a firm (Gomez-Mejia et al., 2007Berrone et al., 2012). Different nonfinancial aspects of the business that satisfy the affective needs of the controlling family members converge in the concept of socioemotional wealth. Five different dimensions have been categorized as the main nonfinancial utilities: family control and influence on the business, identification of family members with the firm, binding social ties, emotional attachment of family members and the renewal of family bonds to the firm through dynastic succession (Berrone et al., 2012). These dimensions may exert different and contrasting effects on a firm's propension to invest in joint-controlled ventures.
Owning family members derive a strong emotional return from family influence and control over the business as they are deeply attached to it (Gomez-Mejia et al., 2007;Zellweger and Astrachan, 2008). The level of proprietary control over the company is itself a source of a strong emotional return for the family and provides it with power and legitimacy in supporting decisions aimed at increasing the nonfinancial returns it derives from the company . The firm is perceived by the owning family as an extension of the family itself because of the tight links and strong sense of identification with the firm (Berrone et al., 2012;Basly and Saunier, 2020). As a matter of fact, there is evidence that the desire to maintain company control firmly in its hands implies that family businesses tend to be more levered than nonfamily firms Bacci et al., 2018) and, consistently, the desire to preserve the sense of identification leads family businesses to avoid opening up their equity to nonfamily investors (Romano et al., 2001). Therefore, the relevance of control and sense of identification for family firms might negatively affect their desire to participate in joint-control ventures in which they would have to share control of the business with nonfamily members.
Family companies are particularly prone to developing social bonds which provide family members with a strong sense of legacy and relevant emotional returns (Berrone et al., 2012;Gomez-Mejia et al., 2011;Astrachan and Jaskiewicz, 2008). Family firms develop strong ties with employees, who become part of a sort of extended family, as well as with suppliers, with the community in which the company operates and with its stakeholders at large (Berrone et al., 2010). Owing to the ability to create social bonds, strictly related to the pursuit of nonfinancial objectives, family companies enjoy a higher reputation than nonfamily firms (Beck and Pr€ ugl, 2018;Gavana et al., 2018;Binz et al., 2013). The relational ability and reputation may help the selection process of the co-venturers and facilitate family firms' jointventure investments.
The SEW dimension related to the emotional attachment of family members highlights the role of emotions in family-business management owing to the mix of shared experiences and events that affect current activities and relationships within a family company (Berrone et al., 2012). Emotions and the relevance of family ties can lead to the selection of family members with limited skills and experience to act as managers (Jaskiewicz et al., 2013) and, therefore, if the company wants to explore new markets, the JV may limit the risks. So, the facilitation of the internationalization processes might be a stimulus for investment in JVs. On the other hand, the presence of family ties and the desire to protect the SEW may lead to executive entrenchment (Gomez-Mejia et al., 2001;Randolph et al., 2018) and induce family members to pursue private benefits at the expense of the firm's stakeholders . Literature points out that family firms are more prone to entrenchment than their nonfamily counterparts and that family owners are abler than managers in widely held corporations to efficiently derive private benefits from the firm (Sharma, 2004;Claessens et al., 2002).
These negative aspects of family firms' governance and behavior may become an obstacle in the search for JV partners. Potential partners may be reluctant to enter into a collaborative relationship with a family business, fearing the negative effects of weak governance or family-centric behavior that could harm them. Families perceive the firm they own as a longterm investment to be passed on to their heirs in order to renew the family bonds to the business and to meet an emotional need to perpetuate the family dynasty (Berrone et al., 2010(Berrone et al., , 2012. For this reason, they receive emotional returns in the present as well as future benefits of control . The preservation of socioemotional wealth affects family companies' risk-taking decision-making in different ways and this is why family businesses are risk-averse and risk-willing at the same time (Gomez-Mejia et al., 2007). There is evidence that family firms are prone to facing the risk of financial performance (Gomez-Mejia et al., 2007) and the risks related to debt financing in order to preserve family control over the business . On the other hand, they avoid entrepreneurial risk and high-outcome variance ventures in order to protect the business' survival (Naldi et al., 2007). Consequently, the desire to renew the family bonds to the firm through dynastic succession produces a long-term time horizon and risk aversion (Berrone et al., 2012;Zellweger et al., 2012) that may affect investments in joint ventures.
The desire to protect the SEW is stronger when the family exerts its influence directly through having a family CEO (Huybrechts et al., 2013). Nevertheless, family firms differ in the extent to which they aim to preserve the SEW depending on the presence of family members on the board , a major source of heterogeneity among family companies (Di eguez-Soto and Mart ınez-Romero, 2019). Family membership represents a "fault line" separating family members from nonfamily members (Minichilli et al., 2010;Sciascia et al., 2013). They might pursue different goals and the dynamics between these groups, as well as the level of family involvement, affect a company's innovation and Family control and influence on JV investment internationalization strategies (Sciascia et al., 2013;Kraiczy et al., 2015; Di eguez-Soto and Mart ınez-Romero, 2019).

Ownership, board characteristics and joint ventures
The nature of a firm, notably family business or nonfamily business, influences the propension to engage in joint ventures when the company chooses to internationalize, or when it pursues innovation goals.
Family firms are characterized by an ability to develop a relevant organizational social capital (Arregle et al., 2007); this results from the firm's interaction and relationships with a wide variety of external stakeholders (Zhara, 2010). Family firms may exploit their unique social capital to identify partners and develop joint ventures in order to innovate (Zhara, 2010). For family firms, collaborative innovation, including joint ventures, may be an effective solution to overcome resource constrains and gain access to partner knowledge (Feranita et al., 2017). On the other hand, the concern to protect the SEW and, in particular, the desire to preserve control over the business, may lower the propensity for family firms to collaborate with external parties (Gomez Mejia et al., 2007;Le Breton-Miller and Miller, 2009). Family firms worry that entering alliances entails a risk of leaking crucial information to the partners and losing a competitive advantage (Zhara, 2005), especially regarding technological collaborations (Nieto et al., 2015). Analyzing new product development process, Cassia et al. (2012) find that family firms tend to collaborate with those partners who are not considered as a threat to the company. In fact, to limit the loss of SEW, when family firms participate in innovation alliances, they prefer partnerships with universities, suppliers and clients rather than with competitors . Concerning internationalization processes, the characteristics of family firms may have opposite effects on the formation of International Joint Ventures (IJVs). Given the risk aversion of family firms, a joint venture is a viable means to limit the risks stemming from international inexperience. In family firms, managers are often hired not for their capabilities but because they are members of the family (Chrisman et al., 2014) and this fact may hinder the internationalization process (Fern andez and Nieto, 2006;Graves and Thomas, 2006). Therefore, inexperienced family firms, compared to inexperienced nonfamily firms, are more likely to choose joint ventures than wholly-owned subsidiaries (Kuo et al., 2012). On the other hand, the tendency of family firms to remain independent from external parties and keep control over the business leads them to prefer wholly-owned subsidiaries to joint ventures (Abdellatif et al., 2010), in particular when the governance quality of the host country is high, diminishing the need to rely on local partners (Chang et al., 2014). Moreover, the strong attachment of family members to the firm negatively affects the strategic sensitivity of family businesses which can lose opportunities to acquire new knowledge from strategic alliances. In such a situation, hiring highly skilled external directors, whose values are not in contrast with those of family members, may help the family to understand that engaging in an IJV may be an opportunity to survive and pass the firm on to future generations (Debellis et al., 2021).
Research has demonstrated that the means of entering foreign markets is affected by the ownership characteristics of the local firm (Sestu and Majocchi, 2020). In fact, family firms have a specific asset which is not tradable on the market: "familiness." When the investing firm is a family firm and the local firm is a nonfamily one, a full acquisition of the local firm is more likely. In this case, a wholly-owned subsidiary allows the investing firm to maintain the family's control and combine its family-specific assets with those of the local firm. On the contrary, when the investing firm and the local firm are both family firms, both their familyspecific assets cannot be separated from the firm. In this case, a joint venture is more likely because it allows both parties to maintain their respective family status and the profits of the joint venture will remunerate the co-venturers for the assets transferred.
Prior studies have highlighted the heterogeneity of family firms, demonstrating how different types of family ownership structure differently influence foreign market entry mode decisions. This is due to the greater, or lesser, importance placed on two main SEW priorities: the long-term survival of the firm and the maintenance of family control (Pongelli et al., 2016). When the level of founder ownership is high, noncooperative entry modes are preferred due to the founder being reluctant to share control and business decisions with external parties. On the contrary, when the ownership structure includes families other than the nuclear family, family firms are more prone to engage in joint ventures rather than rely on wholly owned subsidiaries when they choose equity entry modes.
There is no evidence related to the effect on JV investments of family members on the board, but empirical literature reports that the level of family involvement in management and on the board, respectively, moderate a firm's innovation and internationalization propensity (Di eguez-Soto and Mart ınez-Romero, 2019; Sciascia et al. , 2013).
Based on the theoretical framework and the literary review, we posit the following hypotheses: H1a. Family ownership control negatively affects JV investment size.
H1b. Family involvement on the board amplifies the effect of family control on JV investment.

Link and scale JVs
Family firms may present a different propensity to engage in scale joint ventures rather than in link joint ventures. A scale joint venture is created to achieve economies of scale. The coventurers enter together a contiguous stage of their production or distribution cycle or a new market, bringing similar skills and resources to the joint project. In a link joint venture, the coventurers pursue different objectives and contribute distinctive and complementary skills and resources to the joint project (Hennart, 1988). Since co-venturers are also competitors, a link joint venture gives them the opportunity to acquire partner knowledge and use it to gain a competitive advantage (Grant and Baden-Fuller, 2004;Baughn et al., 1997). Therefore, asymmetric outcomes are more likely to occur in link joint ventures than in scale joint ventures (Dussage et al., 2004). Confirming this risk, the takeover of the joint venture by a partner occurs earlier and more often for link joint ventures than for scale joint ventures (Dussage et al., 2000). As a consequence, given family firms' risk aversion and the differences between link and scale joint ventures and the implications in terms of risk, family businesses might be less prone to invest in link joint ventures than their nonfamily counterparts. H2 is formulated as follows: H2. The type of JV investment moderates the effect of family control on JV investment.

Intellectual capital in family firms
Literature has considered how the nature of the ultimate controlling owner can influence the quality and the development of IC. In this vein, research has highlighted that family firms can rely on peculiar intangible assets that are not accessible to their nonfamily counterparts. Trevinyo-Rodr ıguez and Bontis (2007) have coined the term intellectual capital in family businesses (ICFB), defined as the set of quantitative and qualitative intangible assets that influence a company's performance. In particular, qualitative intangible assets comprise: the founder and/or family's vision, values, energy, social networks, cohesiveness and commitment transmitted to the employees; the processes, procedures and policies of the organization and, finally, the network of the firms that allows the relationship between the different parts of the firm to be maintained and developed. The identification of intangibles Family control and influence on JV investment inherent to family firms has been addressed by Claver-Cort es et al. (2013). They have provided an intellectual capital model which lists the main IC elements specific to family firms, grouped into the three IC categories: human capital, structural capital and relational capital. The literature highlights that human capital represents the most important intangible of family firms (Sirmon and Hitt, 2003). Family firms are characterized by long-term relationships with employees that foster its organizational social capital (Leana and Van Buren, 1999). Claver-Cort es et al. (2015) identify ten human capital intangibles that characterize family firms and find sixty indicators useful to managers in order to measure such intangibles.
Given the peculiarities of family firms, research has also analyzed how the degree of "familiness" (Habbershon and Williams, 1999) affects IC efficiency. By using family ownership and the proportion of family directors on the board to approximate the degree of familiness, Greco et al. (2014) find that, compared to their nonfamily counterparts, family firms are characterized by higher average VAIC TM . Also, the involvement of family members from later generations in the management of the company, as well as the presence of a family CEO, positively affects IC performance in family firms (Ginesti and Ossorio, 2020).
Research has investigated how human, relational and marketing capital affect the innovation capacity of family firms relative to their nonfamily counterparts. The findings suggest a high innovation capacity of family firms because this type of company, compared to nonfamily businesses, shows higher average values for qualified employees and employees engaged in R&D activity (human capital) and outperform nonfamily businesses regarding cooperation agreements (social capital) (Bresciani et al., 2013;Thrassou et al., 2018).
Based on the above discussion we posit the following hypothesis: H3. Intellectual capital efficiency moderates the effect of family control on JV investment.

Data
The study uses data from Italian, Spanish, German and French stock markets for the years 2010-2018 to investigate how family involvement in the firm affects its investment size in joint ventures as well as how other components of a firm's intellectual capital (IC) are related to investments in joint ventures (JVs). Italy, Spain, Germany and France are civil-law countries with similar legal settings, financial market level of development, they are characterized by the prevalence of closely held companies (La Porta et al., 1999;Edwards and Fischer, 1994;Barca, 1995) and by the presence of family businesses that have long-lasting family control (Franks et al., 2012;La Porta et al., 1999;Faccio and Lang, 2002). From the accounting point of view, EU-listed companies adopt the same set of standards for the preparation of their financial statements.
For our empirical investigation, a search was made on the whole population of listed firms in the four countries to identify those with interests in joint ventures in each year from 2010 to 2018. To make up the sample of firms, we relied on JV information provided in the consolidated financial statements downloaded from the firm's web site in the Investor Relation section.
The final sample is composed of all nonfinancial firms with JVs detected in the population of these countries' listed firms. Once companies with missing financial data and missing information on JV type were removed, we had, on average, 132 firms for each year, and we have 855 firm-year observations. We hand-collected the data on joint venture assets from the notes to the co-venturers' consolidated financial statements. All financial statement information and market data for the co-venturers was collected from Orbis, the global Bureau van Dijk database. Ownership data and Board attributes were reconstructed based on the information available on the Orbis database, on the Chamber of Commerce register and on companies' websites.
We defined a family firm as one where a family owns at least 20% of common shares (Villalonga and Amit, 2010). Figure 1 presents the theoretical framework and the research hypotheses developed in this study.

Methods
In order to investigate the effect of family involvement and the moderating role of other components of intellectual capital on JV size, we estimated the following regression model: We estimated model (1) using a GLS panel-data approach with standard errors adjusted for correlation within a cluster, assuming that standard errors are clustered by firm. A panel-data approach uses efficiently the cross-section and time-series data, increasing the parameter's reliability and also reducing the likelihood of multicollinearity problems. To control for industry and year fixed effects, we included dummy variables in the model. I j,i,t is a dummy variable that takes the value of 1 if the co-venturer i is from industry j and 0 otherwise, to control for differences across industries; D t is a dummy variable that takes the value of 1 in year t and 0 otherwise, to allow for differences in the intercept term over time.

Variables
The dependent variable is the size of investment in JV in log terms, measured as the sum of the assets of a firm's JVs (based on the firm's share).
The explanatory and control variables are CEO-D, W-Board, F-CEO, F-Board, HCE, SCE, CEE, Link JV, Size, Roa, Extra-EU Op, year and industry dummies (Table 1).
CEO duality is a dummy variable that takes the value of 1 if the CEO is also the chairperson of the board of directors (Huybrechts et al., 2013). Woman on board is the weight (per cent) of women that sit on the board of directors (Cruz et al., 2019;Nadeem et al., 2017).

Description of variables
Family CEO (F-CEO) is a dummy that takes the value of 1 if the CEO is a member of a family that owns at least 20% of the firm's common shares (Muñoz-Bullon et al., 2018;Ginesti and Ossorio, 2020). Family on board (F-Board) is the weight (per cent) of family members who sit on the board (Sciascia et al., 2013;Greco et al., 2014).
Firm size (Size) is measured as the logarithm of assets (Muñoz-Bullon et al., 2018;Ginesti and Ossorio, 2020). Firms' profitability is proxied by Roa, the ratio of operating income to total assets (Zhara, 2005; Ginesti and Ossorio, 2020). Extra-EU Op is a dummy variable that takes the value of 1 when a firm operates in extra-European Union countries, otherwise it takes the value of 0.
To measure the efficiency of IC components, we use the VAIC methodology developed by Pulic (2004). In Pulic's model, HCE is a proxy for the efficiency of human capital. It is calculated as VA/HC, i.e. value added scaled by total salary and wage duties for the company. Value added is given by VA 5 OUT-IN where OUT 5 total sales and IN 5 cost of bought in materials, components and services.
Structural capital SC is measured as SC 5 VA-HC, and structural capital efficiency SCE is given by SCE 5 SC/VA. Intellectual capital efficiency is then given by ICE 5 HCE þ SCE. Pulic (2004) proposes that the efficiency of capital employed be taken into account as IC cannot create value added on its own: CEE 5 VA/CE where CE is the book value of net assets, so that VAIC 5 ICE þ CEE is a measure of a firm's overall efficiency. Accordingly, we use CEE as a control variable. Since data are obtained from audited financial statements rather than being subjective evaluations, such as questionnaires, VAIC methodology provides a standardized and consistent measure (Shiu, 2006). Iazzolino and Laise (2013) argue that the VAIC approach does not contravene any fundamental accounting principles. In Pulic's work, human capital is defined not as the actual skills, capabilities, etc., of the employees but the amount of investment in employees as a return of the set of employee characteristics and SC is defined not as a set of intangible assets but as the part of VA created by a set of intangible asset characteristics. So, once the concepts are properly comprehended, the VAIC approach "has its own logical coherence" (Iazzolino and Laise, 2013, p. 552).
In order to classify a JV as "link," we followed Laurenco and Curto's methodology. We identified the JVs for each venturer from the notes in the consolidated financial statements. We then collected information about each JV's business, as well as the venturers' business, in order to analyze and draw conclusions about the role of each venturer in the JV. We looked for information about the JV's business and venturers' business on the JV's website as well as on the venturers' websites. When the venturers undertake complementary businesses in different industries, we classified the JV as a link venture, and when the venturers operate in a similar business, the JV was classified as a scale venture. As a co-venturer may participate in more than one JV, in line with Laurenço and Curto (2010), we coded the co-venturer as "link" when the majority (in number) of joint ventures are link. Link JV is a dummy that takes the value of 1 when the majority (in number) of joint ventures of a firm are link, otherwise the dummy takes the value of 0.

Results
We report in Table 2 the breakdown of our data for family and nonfamily firms for the period 2010-2018, showing the mean and standard deviations of the dependent and independent variables. We also present the t-statistics and their p-values to verify the differences between family and nonfamily firms. There were 540 firm-year observations for nonfamily firms and 315 firm-year observations for family firms. Joint venture size is lower for family businesses and the t-statistics indicate that the difference is significant at 1%. Consistently with literature Chen et al., 2005;Lam and Lee, 2008;Bjuggren et al., 2018), CEO Family control and influence on JV investment duality and the percentage of women on the board, as well as human capital efficiency, are significantly different at the 1% significance level between the two samples while there is evidence of difference at the 5% level for the proportion of link-type joint ventures. Table 3 gives the correlation coefficients for the variables used in the regression models for the full sample. The analysis indicates that JV size is negatively and significantly correlated with family CEO, the percentage of family members on the board and Link JV type. The sample presents a positive and highly significant correlation between CEO duality and family CEO, as the overlapping position of CEO and Chair of the board of directors is held by a family member in 82.93% of cases in the family firms subsample. Table 4 presents the generalized least squares (GLS) panel model results, based on the model in Equation (1).
The introduction of family-specific variables and the interaction term confirmed the results obtained for the full sample and the nonfamily subsample. CEO duality always presents a significant negative effect on JV size for family firms whilst it exerts a significant positive effect for nonfamily companies. The percentage of women on the board increases JV size; the effect is significant for all firms and nonfamily firms but never for family businesses. Conversely, Link JV always lowers JV size, but the effect is significant only for family firms. Structural capital efficiency always presents a negative not significant effect on the dependent variable. Human capital efficiency significantly increases JV size only for all firms.
As we have a panel longitudinal dataset, we estimated Equation (1) by calculating clustered standard errors to account for the firm-fixed effects. This procedure gives unbiased standard errors as long as the fixed effects (time and industry) are negligible.
The simultaneity of JV size and VAIC components, and the omission of relevant variables, may also result in endogeneity problems. The direction of causality can be reverse or run both ways. To address these concerns, several approaches can be used: fixed effects, control variables, lagged variables and generalized method of moments (Li, 2016). Following Li (2016), we address these problems with two methods. The first method was to cluster for firm dimension and account for the time and industry-fixed effects with dummy variables. This method tries to control for some of the unobservable determinants of JV size to mitigate the omitted variables problem. The second method was to analyze the relationship between VAIC components on subsequent JV size. The panel GLS regressed JV size in year (t) on CEE, HCE and SCE in year (tÀ1).   Table 1.***Indicate significance at the 1% level; **Indicate significance at the 5% level; *Indicate significance at the 10% level Table 3. Correlations Family control and influence on JV investment Table 5 presents the results using the second method, lagging the CEE, HCE and SCE independent variables and controlling for the year and industrial sector fixed effects.
Both methods give qualitatively similar results in terms of explicative power and significance of the independent variables.
The results of the hypotheses testing are displayed in Table 6.

Family control
Our results show that the overall size of investment in JVs for family businesses is significantly smaller than for nonfamily businesses. This is in accordance with H1a and supports the view that family firms attribute a fundamental value to family control in order to preserve socioemotional wealth, that is the system of nonfinancial returns that the family derives from the business (Berrone et al., 2012;Gomez-Mejia et al., 2011). This result is consistent with empirical studies reporting that family firms are not prone to expanding through cooperative modes, as they are not inclined to share business control and management with nonfamily members (Gomez-Mejia et al., 2007Yamanoi and Asaba, 2018), as well as, in forming innovation alliances, they are concerned with the preservation of their monitoring power Nieto et al., 2015). This result empirically supports the theoretical contribution to family firms' international JVs of Debellis et al. (2021) as they argue that family firms, in spite of their higher ability to establish and govern JVs, have a lower willingness to form them.

Family influence and board characteristics
According to the view that a firm's strategic choices are affected by dominant coalition's background composition (Hambrick and Mason, 1984), and to the stream of literature that has  Table 1.***Indicate significance at the 1% level; **Indicate significance at the 5% level; *Indicate significance at the 10% level Table 4. GLS panel regression results studied family firms' innovation under this perspective (Sciascia et al., 2013;Kraiczy et al., 2014Kraiczy et al., , 2015; Di eguez-Soto and Mart ınez-Romero, 2019), our results highlight the relevance of board characteristics and type of influence exerted by the owning family on JV investments. Consistent with H1b, our results point out that CEO duality reduces JVs investment size in family firms whilst it has the opposite effect in nonfamily businesses. In accordance with previous literature, we find that CEO duality is more common in family-controlled firms than nonfamily firms Chen et al., 2005;Lam and Lee, 2008) and the position is normally held by a family member. CEO duality is the expression of the strong will of the family to maintain business control in order to preserve SEW and it fosters a long-term investment horizon and risk aversion. Literature points out that the concern for the preservation of monitoring power affects family firms' alliance structure Nieto et al., 2015) and that they tend to directly monitor their subsidiaries (G omez-Mej ıa et al., 2010). The overlapping role of CEO and chair of the board denotes a firm's preference for a strong monitoring power that lowers the investment in joint-controlled ventures.
Moreover, CEO duality speeds up the decision-making process but lowers the ability to supervise (Cabrera-Su arez and Mart ın-Santana, 2015). There is evidence that it might exert, Family control and influence on JV investment respectively, negative and positive effects in family and nonfamily companies' performance (Lam and Lee, 2008). In this case, having the chief executive holding the highest position of power on the board of directors might increase the capability to seize joint venture opportunities but it might be perceived by family-business potential co-venturers as the expression of a weak and opaque governance, given their ability to extract private benefits from the business (Sharma, 2004;Claessens et al., 2002), reducing the opportunities to form JVs.
Conversely, a family CEO and the percentage of family members on the board have a negative but not significant effect on JV investment size. This result is consistent with the evidence that family members are less prone to enter cooperative investments in order to preserve the SEW (Gomez-Mejia et al., 2007) and the view that nonfamily directors and nonfamily CEOs rely on more diverse networks (Arregle et al., 2007) that may facilitate the development of joint ventures. Nevertheless, this evidence suggests that family involvement in active management, per se, does not significantly affect a family firm's propensity to invest in joint controlled businesses, but the effect depends on CEO power.
Not having found a significant effect for the percentage of family members on the board might also depend on the different generational stages the observed firms belong to. In earlier generational stages, family members on the board belong to the same nuclear family, the influence of the founder is stronger, as is the long-term perspective, so a strong family presence on the board leads to a preference for investments that are under the family's complete control (Gomez-Mejia et al., 2007). In later generational stages, family members on the board might represent different generations as well as belong to different family branches (Le Breton-Miller and Miller, 2013); they may prefer to pursue financial objectives compared to SEW (Mariotti et al., 2020) and have a different risk propension (Muñoz-Bullon et al., 2018) and a shorter term perspective that might increase investment size in JVs. This interpretation is supported by literature reporting that firms are less prone to engage in international JVs when ownership is concentrated in the hands of the founder and more prone when ownership is held by several members not belonging to the same nuclear family (Pongelli et al., 2016). The former ownership structure characterizes the first generational stage and the latter the later ones.
Our results highlight that, in nonfamily companies, the presence of women on the board positively affects JV investment size. This evidence is consistent with research on gender diversity reporting that women's representation on the board positively affects a firm's commitment toward its stakeholders (Nadeem et al., 2017) as well as its reputation (Bear et al., 2010) and this might facilitate the development of its relational capital and the creation of JVs.
The relation between gender diversity and JV investment size is not significant for family firms, and this evidence could be the result of mixed effects. As a matter of fact, female directors may be family or nonfamily affiliated and, in the majority of gender-diverse boards, at least one woman belonged to the first type (Bianco et al., 2015). Family and nonfamily women on the board may have a different effect on a firm's JVs investment size as the former are strongly led by SEW preservation and the latter by financial goals. Empirical evidence on the effect of family and nonfamily female directors on family businesses' CSR engagement supports this interpretation .
Moreover, empirical literature points out that attaining women critical mass on the board enhances the level of firm innovation, but the relation is mediated by board members' strategic tasks (Torchia et al., 2011). Under this perspective, our results call for further research in this direction as they might suggest that the degree to which board members are involved in the strategic process could be different in family and nonfamily firms resulting in a significant positive effect of female directors on JVs investments only for the latter.

JV type and IC components
Our results point out that family firms are more likely to implement link-type JVs than nonfamily companies. This is probably because link JVs are an alternative means of entering foreign markets to a wholly owned subsidiary. This result is in line with recent research pointing out that family firms, compared to nonfamily businesses, are more likely to choose JVs rather than WOSs when entering a new market, due to their risk aversion, as they would reduce the risks linked to the inexperience of their management in international markets (Kuo et al., 2012) and rely on more strategic flexibility (Hitt et al., 1998;Harrison et al., 2001). This explanation is supported by the evidence that a firm's extra-European Union operativity has a significant positive effect on the investment in JVs for family firms.
Nevertheless, consistent with H2, we can note that, in general, family firms tend to limit the size of their investment in JVs when there is a predominance of the link type. The level of risk is different in the case of scale and link JVs. A scale JV is a homogeneous cooperation and "the venturer's relationship is similar to that with a third party with whom it has an arm's-length contract, but where the venturer is protected by a guarantee, that is the common control over that third-party" (Laurenço and Curto, 2010, p. 740). A link JV is a heterogeneous cooperation as it is often created when the venturers come from different industries to enter a new business together and each provides a different and specific contribution in order to develop the new business. In this case, the JV substitutes a contract between the venturers. According to Dussage et al. (2000), partners are more likely to reorganize or takeover link JVs than scale partnerships, whilst scale JVs are more stable and tend to go on without material changes. A link JV is riskier than a scale JV, so, in cases of link JVs, family firms might limit their investment size in order to limit the related risk. This effect is significant only for family firms. They are more risk averse than nonfamily firms as they are concerned for the business' survival over generations: the firm is a long-term investment that must be passed on to heirs and, for this reason, families are more prone to limit the size of investments in link type JVs.
As for the effect of IC, our results point out that structural capital efficiency has a negative not significant effect on JV investment size, in spite of its potential role in facilitating the management of a joint venture's different aspectspartner selection, negotiation, contract formulation and relationship terminationby means of well-constructed organizational structures and in storing and spreading through the organization knowledge acquired by alliances (Chang et al., 2008).
Our empirical evidence shows that HCE is higher for family firms and that the difference is statistically significant. This evidence is consistent with literature highlighting that human capital represents the most important intangible of family firms (Sirmon and Hitt, 2003). Human capital efficiency significantly affects the investment in JVs for all firms, but we do not find this effect for family firms, not confirming H3. A company's resource profile is a relevant component of the alliance formation process (Stuart, 2000); it also affects a firm's ability in accumulating and sharing through the organization what is learnt in previous cooperative agreements, also to the benefit of future alliances as reported by research on nonequity strategic alliances (Chang et al., 2008). Our results suggest that this holds also for nonequity alliances and a specific resource, human capital, has a key role in the development of JV investment size. Nevertheless, we do not find this effect for family businesses. These results, all in all, suggest that if a family firm can rely on IC efficiency it might be less prone to develop joint ventures in order to protect its competitive advantage and, by this means, the SEW. This view is consistent with previous research pointing out family business concerns for the effect of alliances on the preservation of competitive advantage (Zhara, 2005).

Concluding remarks: contributions, implications and limitations
This study investigates the effect of the main dimension of socioemotional wealthfamily control and influenceon a firm's JV investment size, controlling for the effect of JV type and Family control and influence on JV investment of the interplay with the other IC components. It highlights the need to consider family businesses as heterogeneous groups (Chrisman et al., 2005;Westhead and Howorth, 2007;Kraiczy et al., 2015) by taking into account different board characteristics and by integrating IC components and JV type, contextual factors of heterogeneity that help to clarify family firms' investment choices in JVs. In doing so, it also contributes to the stream of literature which analyzes the influence of family firms' board characteristics on a company's innovation strategies (Kraiczy et al., 2014(Kraiczy et al., , 2015; Di eguez-Soto and Mart ınez-Romero, 2019), providing evidence that strong direct influence of the family on the board, by the means of CEO duality, reduces the investment size in JVs. This implies that the presence of family members on the board does not have per se a significant moderating effect on innovation strategies via cooperation, whilst the overlapping role of CEO and Chair of the board may produce this effect. These results have implications for family firms and practitioners as they point out how the choices in terms of board structure may limit the development of JVs, a viable way of acquiring information, knowledge and technology and avoiding the effect of conservatism on business management.
This study also adds to the stream of literature that addresses the issue of gender diversity in family business boards (Martinez Jimenez, 2009;Campopiano et al., 2019;Cruz et al., 2019) pointing out that the presence of women on the board increases JV investment size in nonfamily firms, but this effect is not significant for family firms. This result has relevant implications for family firms' governance and management as it suggests that family businesses should better exploit women's skill, experiences and useful networks in order to develop alliances and increase women's presence when the family exerts a strong direct influence on the board.
This research contributes to JV literature by pointing out that JV investment size is lower for family-owned businesses than for nonfamily firms. Moreover, to the best of our knowledge this is the first study to provide evidence on the moderating effect of JV type and IC components on JV investment, with further implications for management. Family businesses are more likely to invest in link JV than nonfamily companies, but the former tend to reduce the risks associated with this type of alliance by limiting the size of the investment. These results make family businesses aware of their different behavior compared to their nonfamily counterparts, as in preserving the SEW they tend to limit a critical external knowledge source. As a matter of fact in link JV each co-venturer contributes distinctive knowledge, but at the same time the investment in this type of business triggers the development of new capabilities. Human capital has a significant positive effect on JV investments for all firms, suggesting that investing in the former and increasing its efficiency can improve the ability to develop alliances. Our empirical evidence points out that family firms present a higher human capital efficiency than nonfamily companies and they might exploit it in order to expand their joint controlled investments. This result is also of interest as human capital efficiency is a prerequisite to ensure that organization effectively absorbs the knowledge, which could be acquired by the means of JVs (Chang et al., 2008).
In common with applied research more generally, this study is not without limitations. Improvements might be carried out following three different lines: (1) Methodology (2) JV characteristics analyzed (3) Board of director dimensions considered The study focuses on the effect of "family control and influence" on the investment in JVs. Other SEW dimensions, such as family members' sense of identification with the company and the desire to transfer the business to future generations, might affect JV investments. We cannot rely on this type of information as we based our analyses on data contained in the financial statements and in the Orbis database. It can be codified by integrating the study with information collected through surveys. In doing so, future research could go further in order to develop a multidimensional measurement of family-affective endowment in the firm.
The research model includes, for the first time, the type of alliancelink or scaleas a variable expected to moderate the effect of family control on JV investment size. As a matter of fact, other contextual factors, such as governance characteristics of the co-venturers, distance and institutional aspects of the country in which the JV operates might affect family firms' risk perception. This could be an interesting line of investigation for further research.
This explorative study analyses the effect of certain board characteristics (CEO nature, CEO duality, proportion of family members and women) on JV investment but other individual characteristics of board membersexperience, values and personalities - (Hambrick and Mason, 1984) could help explain a firm's behavior. Our results call for a more detailed analysis on the effect of board diversity in order to better disentangle the effect of family firms' heterogeneity on joint-controlled investments. So, the next step in this research will add a further dimension of observation taking into account the internal composition of the family group on the board in terms of generations, family branches sitting on the board and ownership stake held by board members. This additional dimension will allow us to identify potential dysfunctional group dynamics that can limit family firms' investments in JVs.