INTRODUCTION

In the last two decades, strategy and international business scholars have been increasingly interested in studying family-controlled firms (FCFs). FCFs are enterprises that are “governed and/or managed with the intention to shape and pursue the vision of the business held by a dominant coalition controlled by members of the same family or a small number of families in a manner that is potentially sustainable across generations of the family or families” (Chua, Chrisman, & Sharma, 1999: 25). The most important findings in this prior research are: the FCF is the most common organizational form in the majority of national contexts (Burkart, Panunzi, & Shleifer, 2003); FCFs follow distinct strategies that differentiate them from non-FCFs, such as long-term orientation, risk avoidance, preference for equity financing, and commitment to frugal innovation (Le Breton-Miller & Miller, 2006; Sirmon & Hitt, 2003); and, possibly due to their strategic distinctiveness, FCFs tend to outperform non-FCFs (Anderson & Reeb, 2003; Villalonga & Amit, 2006). In sum, this literature has produced a coherent narrative about family firms, centered on the ‘stylized facts’ of family firm ubiquity, strategic uniqueness, and financial outperformance.

Upon closer inspection, however, all of these ‘stylized facts’ can be challenged. For example, FCFs represent well over 50% of all publicly listed firms in countries such as Chile, France, India, and South Korea (Duran & Ortiz, 2020; Jameson, Prevost, & Puthenpurackal, 2014; Kim, Kim, & Lee, 2008; Sraer & Thesmar, 2007) but only a quarter or less in Austria, China, and the UK (Amit, Ding, Villalonga, & Zhang, 2015; Faccio & Lang, 2002; Krivogorsky & Burton, 2012). Furthermore, motivated by stewardly convictions (Le Breton-Miller & Miller, 2009) or socioemotional wealth preservation (Gómez-Mejía, Cruz, Berrone, & DeCastro, 2011), FCFs make distinctive strategic decisions in certain countries, yet display isomorphic tendencies elsewhere (Arregle, Hitt, Sirmon, & Very, 2007). Finally, whereas FCFs outperform non-FCFs in India, Japan, the US, and Poland (Anderson & Reeb, 2003; Kowalewski, Shaefer, Stetsyuk, & Talavera, 2010; Sarkar & Sarkar, 2000; Yoshikawa & Rasheed, 2010), they underperform in Egypt, Peru, and New Zealand (Boone, Colombage, & Gunasekarage, 2011; Elsayed, 2011; Prabowo & Simpson, 2011). Do these conflicting findings point to an intellectual crisis in the field of family firm scholarship (cf. Gedajlovic, Carney, Chrisman, & Kellermanns, 2012), or is there a theoretically satisfactory explanation for these contradictions?

Drawing from comparative institutionalism, the literature has provided formal institutional arguments to explain cross-country differences in FCFs’ prevalence, strategy, and performance, suggesting that legal and regulatory institutions (La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 1998; Peng & Jiang, 2010) and institutional imperfections or “voids” (Khanna & Yafeh, 2007; Luo & Chung, 2013) are important contingencies affecting FCFs’ ownership and control, unique strategic patterns, and performance (dis)advantages. While recognizing the importance of formal institutions, we propose that the seemingly contradictory empirical findings in the FCF literature also derive, at least in part, from previously overlooked country-level informal institutions. Thus, the research question addressed in this study is: How do national informal institutional environments in which FCFs are embedded affect the prevalence, strategic choices, and financial performance of these firms?

Our main thesis is that a country’s informal institutions affect the constitutive legitimacy of FCFs (Rao, 2002), which in turn influences the prevalence of family ownership and the degree to which family firms can leverage their distinctive advantages for unique strategies and performance outcomes. Constitutive legitimacy encompasses both the degree of passive taken-for-grantedness (Suchman, 1995) and active socio-political endorsement (Aldrich & Fiol, 1994; Li, Yang, & Yue, 2007) FCFs can muster as an organizational form. To capture FCFs’ constitutive legitimacy, we introduce the concept of “family business legitimacy” (FBL). FBL reflects the degree to which a country’s environment is characterized by a set of social ordering systems, social relationships, and values that recognize the family firm as the basic unit of economic production, and kinship ties – as the predominant conduit of social and economic exchange. FBL is captured by a subset of a country’s informal institutions, which impact FCFs’ pervasiveness, strategic choices, and performance, apart from and beyond relevant formal institutions. To assess FBL empirically, we develop a new country-level index, the “family business legitimacy index” (FBLI), which provides a fine-grained assessment of informal societal institutions that are particularly relevant for FCFs. The measurement instrument consists of 20 items grouped into five formative dimensions: intergenerational survival orientation, continuity orientation, network-based relations, in-group solidarity, and patriarchal orientation. Using the index, we map the cross-national variance of FCFs-relevant informal institutions in 83 developed and emerging economies. Our index is orthogonal to indicators measuring formal institutions that are commonly theorized to support or constrain FCFs.

Our study employs advanced meta-analytic and censored regression techniques, allowing us to analyze and synthesize results reported in 484 primary studies. We find support for our main proposition, showing that FCFs’ characteristics vary by country and that the informal institutional environment in which FCFs are embedded significantly impacts their prevalence, strategic choices, and performance. These findings are econometrically robust, even when controlling for formal legal and financial institutions that have been known to impact FCFs behavior and outcomes (La Porta et al., 1998). Together with the development of the FBLI index, this analysis constitutes an important empirical contribution, as it helps explain and reconcile conflicting findings in extant research.

We seek to make some theoretical contributions as well. More broadly, we add to the literature on institutional embeddedness of organizations by showing that informal institutions, in addition to formal institutions, affect organizational forms as a whole (Cantwell, Dunning, & Lundan, 2009; Lu, Song, & Shan, 2018; Sartor & Beamish, 2014). More specifically, we inform family firm research suggesting that family control emerges at the intersection of both formal and informal institutional systems, with informal institutions supporting or constraining the pursuit of family-centric goals within the range of possibilities offered by the formal institutional framework. In this sense, we depart from prior studies that have viewed informal institutions as a substitute for weak formal institutions (cf. La Porta et al., 1998; Peng, Sun, Vlas, Minichilli, & Corbetta, 2018). Scholars should, therefore, more explicitly consider informal institutions as a separate and independent influence on firm behavior. This is an important insight adding to previous comparative work, in which informal institutions have rarely been studied (cf. Chrisman & Patel, 2012; Miller, Le Breton-Miller, & Lester, 2012). The construct of FBL introduced here, and the proposed FBL index to measure it, should open the door for much-needed future research in this area.

THEORY AND HYPOTHESES

Institutional Embeddedness of Organizations

All countries have distinct institutions, often referred to as the ‘rules of the game’ in society (North, 1990; Scott, 2001). Variance in terms of both formal and informal institutions creates different contexts to which businesses must adapt (Kostova, 1999). The manifestations of such adaptation frequently involve the adoption of organizational forms, practices, or procedures, which are deemed desirable by the organization’s external constituencies (DiMaggio & Powell, 1983; Meyer & Rowan, 1977; Tolbert & Zucker, 1983). Formal institutions are typically codified in regulations, laws, and rules; they affect production and transaction costs, and through that, the profitability of economic activities (North, 1990). Informal institutions too “codify” appropriate practices for organizations and affect economic transactions, but they derive from networks of social relationships, power elites, and hierarchies of status and domination. They are self-enforcing, because violating them could render the organization illegitimate (Guler, Guillén, & Macpherson, 2002; Kostova & Zaheer, 1999). Prior research has explained cross-country variation in FCFs’ prevalence, behavior, and outcomes primarily based on formal institutional embeddedness (e.g., in the legal and regulatory environment) (La Porta et al., 1998; Peng & Jiang, 2010; Peng et al., 2018). We complement this view by examining the impact of informal institutions on organizational forms like FCFs.

National informal institutional environments are reflected in three broad and interrelated elements (cf. Friedland & Alford, 1991). First, an informal social order system that captures how institutions and individuals have organized themselves to overcome collective action problems and assign social groups to semi-permanent class structures and social categories (Cooley, 1992; Giddens, 1987). Second, social relationships that structure and stabilize the dynamics between organizations and people and establish a basis upon which actors can mobilize resources and participate in social exchanges (Portes, 1998; Tsai & Ghoshal, 1998). Third, values or guiding principles that define the cultural identity of a society as a whole and that guide societal behavior by shaping perceptions of reality (Harrison & Huntington, 2000; Rokeach, 1973). These three elements are interrelated and mutually reinforcing. Social order systems, such as class structures, influence social relationship formation due to, for example, homophilic tendencies in social network evolution (McPherson, Smith-Lovin, & Cook, 2001). Social relationships that exist within the institution of the family but also within paternalistic organizational forms like FCFs shape individuals’ values through primary and secondary socialization (Chao, O’Leary-Kelly, Wolf, Klein, & Gardner, 1994). Finally, values tend to reinforce social order systems, because they both strengthen social cohesion in “in-groups” and serve as a normative platform supporting collective action and institutional maintenance (Dacin, Munir, & Tracey, 2010).

These three elements jointly determine which organizational forms are seen as ‘constitutively legitimate’ (Rao, 2002) and which are looked at with suspicion. Firms that are perceived as legitimate are accepted, understood, and actively supported by other actors in their organizational field, including customers, suppliers, employees, investors, and governments (Carroll & Hannan, 2000; Zucker, 1983). Constitutive legitimacy is granted when firms adopt legitimate organizational forms and behave in socially acceptable ways. In contrast, adopting an organizational form that is less consistent with societal preferences can negatively impact their functioning, survival, and success (Dobrev & Gotsopoulos, 2010).

Family Business Legitimacy as a Contingency Factor

The concept of family business legitimacy (FBL) can be seen as a property of the informal institutional context, which derives from the social order systems, social relationships, and values in a given country. FBL is an outcome of what Davis (2005) refers to as “a set of structures, processes, and institutions … around organizations that allocate power and resource control amongst participants” (p. 143). In strong FBL countries, the family is positioned as the central economic unit, kinship-based forms of social exchange are favored, and the business culture is congruent with typical family values. Social order systems support informal hierarchical arrangements and well-defined status and authority patterns (Carney, 2005). Relationship systems, finally, are centered on kinship, patriarchy, and long-standing social ties (Khatri, Tsang, & Begley, 2005), and on cultural values like deference, reciprocity, and saving face (Bertrand & Schoar, 2006). Such contexts favor leaders who adhere to patriarchal traditions and foster in-group collectivism (Howell et al., 2007). Senior managers are viewed as the patriarchal heads of the family and their authority is expressed in paternalistic terms, with family CEOs staying for a long time at the helm of the firm, even when the firm is not doing well (Gómez-Mejía, Nuñez-Nickel, & Gutierrez, 2001). Since FCFs are generally characterized by the same features, in such contexts, they are likely to be perceived as constitutively legitimate and as appropriate vehicles for controlling private wealth and pursuing entrepreneurial opportunities.

Our central thesis is that FBL is a salient institutional condition that determines the susceptibility of national contexts to family enterprise and thus can explain the cross-country variation in FCF pervasiveness, enacted strategies, and financial performance. The next three sub-sections of the paper present our reasoning and Fig. 1 graphically depicts our propositions.

Figure 1
figure 1

A model of family business legitimacy on family-controlled firm prevalence, strategic differentiation, and performance.

Full size image

Family Business Legitimacy and Family Firm Prevalence

We predict FCFs to be more prevalent in strong FBL countries. In any social setting, managers and entrepreneurs choose ownership and governance structures that align with the expectations of resource-controlling actors. Therefore, in societies that prefer economic transactions to be organized along family lines (that is, in high FBL countries), firms are likely to follow a family-based ownership and governance structure (Greenwood, Raynard, Kodeih, Micelotta, & Lounsbury, 2011). Moreover, in a strong FBL country, FCFs are likely to obtain a taken-for-granted status, which in turn sets the stage for the founding of other similar organizations, thus increasing their prevalence (Carroll & Hannan, 1989). FCFs are congruent with the core traits of stronger FBL contexts. They have family-centered and informal hierarchies and are structured and administered in particularly gendered ways, resembling mini-patriarchies. Decision-making in FCFs is highly centralized, with a minimal delegation of authority and responsibility (Kiong, 2005). The dominant leadership style, often deeply rooted in the nature of familial relationships in society (Howell et al., 2007), is paternalistic. Family owners often govern FCFs like they control their families – creating organizational arrangements that mirror the moral authority of the male head of the household (Weidenbaum, 1996), assuming multiple roles as a way of exerting maximum formal and informal control (Mustakallio, Autio, & Zahra, 2002), and creating an environment where employees expect job security and being looked after by their managers as people, not merely employees (Aycan, Schyns, Sun, Felfe, & Saher, 2013). Societal support for such paternalistic ownership and management patterns is not universal; it is specific to FBL-oriented countries and is largely absent in weaker FBL contexts.

FCFs are also congruent with FBL-oriented contexts because they redistribute economic and social resources along family lines. Whereas FCF principals sometimes ignore pressures from non-family stakeholders, they tend to be more receptive to family members (Gómez-Mejía, Larraza-Kintana, & Makri, 2003). Moreover, FCFs enjoy exclusive social benefits arising in closed kinship-based networks, including social capital, relational trust, and feelings of interpersonal solidarity (Chua, Morris, & Ingram, 2009). Such relational contracting is common in stronger FBL societies (James, 2008) where business leaders often prioritize trust and friendship over costs and profits (Howell et al., 2007). Finally, FCF owners use their business as a vehicle to propagate particular family values (Handler, 1990), which become their organizational culture (Gómez-Mejía et al., 2011). We contend that societal approval of such nepotistic redistributive practices and dominance of family values is not universal; it is common in FBL-oriented countries but largely absent in weaker FBL contexts.

In short, we expect FCFs to be the preferred organizational form in stronger FBL contexts. Their alignment with institutional expectations will make them more legitimate in the eyes of societal stakeholders (cf. Li et al., 2007; Luo, Jeong, & Chung, 2019). Hence, more economic activities will likely be channeled through FCFs, as compared to other organizational forms. Therefore, we propose:

Hypothesis 1:

The stronger the family business legitimacy (FBL) in a country, the greater the prevalence of family-controlled firms (FCFs).

Family Business Legitimacy and Family Firm Strategy

Unlike other firms that tend to follow standard, often globally established strategies, FCFs prefer strategies uniquely favoring family businesses (e.g., preserving socio-emotional wealth; Gómez-Mejía et al., 2007, 2011). Such choices, however, could be seen as a deviation from the norm and be sanctioned by the legitimating environment. We expect that in FBL-oriented countries, FCFs will have more leeway to pursue such unique strategies because of their constitutive legitimacy. Governance scholars have examined a similar phenomenon of adoption of practices that do not conform to the dominant governance logic in a particular country, which they refer to as ‘corporate governance deviance’ (Aguilera, Judge, & Terjesen, 2018). Following the same logic, we argue that in high FBL countries, FCFs are likely to be given greater freedom to choose unique strategies, pursue self-selected objectives, and resist homogenization relative to non-FCFs (Boutilier, 2009; Luo et al., 2019).

Due to their constitutive legitimacy, we expect FCFs in FBL-oriented countries to show greater strategic differentiation and pursue FCF-specific objectives–prioritizing socioemotional wealth over financial gain (Berrone, Cruz, Gómez-Mejía, & Larraza-Kintana, 2010), adopting a parental, altruistic leadership style (Schulze, Lubatkin, & Dino, 2003), choosing strategic alternatives that require patient capital (Arregle, Duran, Hitt, & van Essen, 2017), and reinforce family control (Chirico, Gómez-Mejía, Hellerstedt, Withers, & Nordqvist, 2019). In high FBL countries, such departures from the strategies of dispersed ownership firms is tolerated (Barnett, Eddleston, & Kellermanns, 2009). Because societal actors tend to identify more with FCFs (Astrachan, Klein, & Smyrnios, 2002), idiosyncratic FCF-specific practices and ideas diffuse more readily (Ansari, Fiss, & Zajac, 2010).

The differentiation of FCFs’ strategies from non-FCFs’ strategies will be greater in stronger FBL contexts, especially on dimensions that harbor the risk of surrendering familial control (Chua et al., 1999). These include less diversification (Gómez-Mejía, Makri, & Larraza-Kintana, 2010), less leverage (Mishra & McConaughy, 1999), lower innovation (Duran, Kammerlander, van Essen, & Zellweger, 2016), less internationalization (Arregle et al., 2017), and lower risk-taking in general (Bennedsen, Pérez-González, & Wolfenzon, 2010). These differentiations protecting the affective endowment, or the socioemotional wealth families derive from continued control of the firm (Berrone, Cruz, & Gómez-Mejía, 2012), are likely to persist for two reasons. First, non-FCFs are not likely to duplicate strategies that serve FCF-specific goals because they are less concerned with the preservation of socioemotional wealth. Second, the aforementioned strategic deviations may require complementary assets like social capital mustered by the family, reputational capital contained in the family name, and ‘patient’ financial capital. All of these resources are beyond the reach of non-FCFs, especially in strong FBL contexts.

In contrast, in weak FBL contexts, adopting practices similar to those of non-FCFs might be the only way to achieve legitimacy for FCFs (DiMaggio & Powell, 1983; Miller et al., 2012). Due to the higher stakeholder scrutiny in such contexts, FCFs are forced to blend in with non-FCFs in terms of their strategic make-up (Miller et al., 2012). Hence, in such settings, there will be no significant differences between the strategic choices made by FCFs and non-FCFs on the aforementioned dimensions of diversification, leverage, innovation, internationalization, and risk. Thus, we propose:

Hypothesis 2:

The stronger the family business legitimacy (FBL) in a country, the greater the differentiation between FCFs’ and non-FCFs’ strategic choices in the areas of diversification, leverage, innovation, internationalization, and risk.

Family Business Legitimacy and Family Firm Performance

When organizations receive positive social evaluations, they tend to outperform less-positively evaluated rivals in terms of both accounting profits and stock market valuation (Heugens & Lander, 2009). Positive social evaluations come in different forms, including organizational status – superior standing in an intersubjective, esteem-based ranking system (Podolny, 1994), reputation – favorable generalized expectations about future behavior and performance (Rindova, Pollock, & Hayward, 2006), and legitimacy – cognitive or normative endorsement of a company based on its alignment with prevailing cultural norms, rules, and laws (Scott, 2001; Suchman, 1995).

We expect FCFs to have performance advantages over non-FCFs, especially in FBL-oriented countries, due to both non-market and market forces (Baron, 1995). With regard to non-market factors, in high FBL societies, FCFs will find it easier to receive the endorsement of politicians, government officials, and community leaders (Duran, Kostova, & van Essen, 2017). Such positive relationships provide symbolic support for their actions as well as access to material resources and technical information that are exclusively distributed through concessions and government mandates (Chung & Luo, 2013). With regard to market factors, in FBL-oriented countries, FCFs can develop relational contracts with clients, financiers, and employees, all of whom tend to favor doing business with those types of firms (Lester & Cannella, 2006; Luo et al., 2019). Such social capital residing with the family members associated with the FCF (Arregle et al., 2007) facilitates access to resources that are not accessible to non-FCFs. Furthermore, the long-term perspective of FCFs enables them to develop stable relationships with banks, suppliers, and clients, allowing them to capitalize on their unique strategic characteristics (Le Breton-Miller & Miller, 2006). Finally, the community-level social capital emanating from FCFs’ interactions with a broad range of external stakeholders (Lester & Cannella, 2006) provides access to information necessary to identify market opportunities and create new businesses, which help them develop future revenue streams, generate profits, and realize growth (Zahra, 2010). Therefore, in FBL-oriented contexts, family control will have organizational and relational advantages, which in turn will lead to better financial performance compared to non-FCFs. Therefore, we propose:

Hypothesis 3:

The stronger the family business legitimacy (FBL) in a country, the greater the performance advantages of FCFs relative to non-FCFs.

METHODOLOGY

Sample and Coding

We use advanced meta-analytic and censored regression techniques to test our hypotheses. Our meta-analysis is consistent with recent applications and guidelines (e.g., Buckley, Devinney, & Tang, 2014; Cao, Jayaram, Liu, & Lumineau, 2018; Gonzalez-Mulé & Aguinis, 2018; Stanley et al., 2013). We built the dataset for the meta-analysis in five steps: (1) we read reviews (Amit & Villalonga, 2014; Gómez-Mejía et al., 2011; Sharma, Chrisman, & Gersick, 2012) and meta-analyses (O’Boyle, Pollack, & Rutherford, 2012; van Essen, Carney, Gedajlovic, & Heugens, 2015; Wagner, Block, Miller, & Schwens, 2015) on the topic of FCFs; (2) we combined search terms like business, control, families, family, firm, founder, founding, lone, and ownership, to identify FCF studies in electronic databases (Google Scholar, ABI/INFORM Global, JSTOR, EconLit, and SSRN); (3) we manually searched journals that often publish articles on FCFs; (4) we performed a two-way “snowballing” technique, back-tracing the references in and forward-tracing citations to the previously retrieved articles; and (5) contacted authors to ask for unpublished works or correlation tables in case manuscripts did not report this information.

We followed three exclusion criteria to determine the relevance of retrieved studies for our research purposes (APA, 2008). First, to ensure comparability, we only included studies reporting results on publicly listed FCFs. We excluded studies dealing with private FCFs only or reporting results from mixed samples without differentiating between private and public firms. Second, we only included studies that compared FCFs’ strategic choices and performance with that of a control group of non-FCFs. Third, we removed studies relying on overlapping samples (Wood, 2008). We obtained a final sample of 484 primary studies (309 published papers, 153 unpublished manuscripts, and 22 theses), spanning 83 countries and covering the 1953–2011 time period (see Appendix A).1 We then read all studies and developed a coding protocol (Lipsey & Wilson, 2001). To ensure coding reliability, one author coded all primary studies and another coded a random sub-group of effect sizes (Stanley et al., 2013). The inter-rater agreement (Cohen’s kappa) was 0.98 (Cohen, 1960).

Family Business Legitimacy Index

Following previous efforts to develop formative indexes for measuring variables in strategic management (Crossland & Hambrick, 2011; Martynova & Renneboog, 2008) and international business (Joshi & Lahiri, 2015; Zanakis, Newburry, & Taras, 2016), we developed a formative index to capture countries’ family business legitimacy, the family business legitimacy index (FBLI). Unlike reflective scales, in which the items exhibit properties of a latent construct, the items of a formative scale are jointly constitutive of the construct. Compared to reflective measurement, formative measurement more accurately captures the direction of causality between construct and indicators (Diamantopoulos, Riefler, & Roth, 2008). We followed the five-step formative index development procedure suggested by Diamantopoulos and Winklhofer (2001) – scoping the index and generating items based on the conceptual definition of the construct, followed by item reduction through content validation, multicollinearity analysis, and external validation. Appendix B provides the details of the instrument development procedure. The FBLI is composed of 20 items grouped in five dimensions: (1) intergenerational survival orientation capturing an “implicit contract between generations” (Rosenzweig & Wolpin, 1985: 961) to promote FCF longevity; (2) continuity orientation indicated by a “hidebound attachment to the past” (Miller, Steier, & Le Breton-Miller, 2003: 513) in the administration of FCFs; (3) network-based relations, which encompasses the “community-level social capital generated by the network” (Lester & Canella, 2006: 755) of FCFs; (4) in-group solidarity, which reflects the primacy of FCFs as a vehicle for economic exchange in societies in which “the average radius of trust of cooperative groups tends to be small” (Fukuyama, 2002: 32); and (5) patriarchal domination, which describes how FCFs help sustain “conspicuous gender disparities in the distribution of work and reward [through the] (…) political construction of the family/firm head” (Greenhalgh, 1994: 746).

Table 1 shows the FBLI scores for the 83 countries in our sample, including the normalized FBLI score and the scores for each of the five underlying dimensions. Appendix C describes the breakdown of the 20 items over the five dimensions, the weight of each dimension in the index, and the source of each item.

Table 1 Family business legitimacy and formal institutions indexes per country
Full size table

Family-Controlled Firm Prevalence: Tobit Regression

Hypothesis 1 addresses FCF prevalence. Testing it requires a criterion to separate FCFs from non-FCFs. As the literature has not yet converged on a single definition of FCFs (Duran et al., 2016), we conduct our analyses using the three most common definitions: (1) ownership definition: firms with a significant degree of family ownership, measured either by voting or cash flow rights (Peng & Jiang, 2010); (2) management definition: firms with a significant presence of family members in top management positions (Miller, Le Breton-Miller, Lester, & Canella, 2007); and (3) ownership and/or management definition: firms with a significant family presence in ownership and/or top management positions (Anderson & Reeb, 2003; Gómez-Mejía et al., 2003).

To test Hypothesis 1, we performed Tobit regressions. Tobit regression is intended for continuous data that are censored or bounded at a limiting value (Tobin, 1958). It uses the maximum likelihood regression estimator to avoid a deflated estimation of the slope coefficient and an inflated estimation of the intercept that would result from the use of the ordinary least squares regression estimator on such data (Amemiya, 1973). Our study warrants Tobit regression since our observed dependent variable is continuous with a floor of 0% and a ceiling of 100% (Long, 1997). We control for formal institutions, which is through a composite index capturing the extent to which governmentally enforced initiatives in a given country are aligned to support a pro-market environment (Dau, 2013). It consists of five variables: (1) non-French origins of the legal system (La Porta et al., 1998), a dummy variable equal to one if the legal origin of the country’s law system is non-French (i.e., English, German, or Scandinavian), and zero otherwise; (2) anti-self-dealing index (Djankov, La Porta, Lopez-de-Silanes, & Shleifer, 2008), which measures the legal protection of minority shareholders against expropriation by corporate insiders; (3) creditor rights (Djankov, McLiesh, & Shleifer, 2007), a measure that proxies for protection and rights of secured lenders; (4) minority shareholder rights (Guillen & Capron, 2016), which vets the level of protection minority shareholders enjoy “against the action of large shareholders and/or management and in the event of a change in corporate control” (Guillen & Capron, 2016: 136); and (5) judicial efficiency (La Porta et al., 1998), measuring the level of judicial quality and enforcement in a country. Table 1 shows the normalized score of formal institutions per country. As expected, the FBLI measure is negatively correlated with the formal institutions variable included in our empirical models (r = − 0.37).2 We also control for stock market capitalization to GDP (World Bank) since FCFs are more prevalent in contexts with undercapitalized financial markets (Burkart et al., 2003). Finally, we control for the natural log of GDP per capita.

Family-Controlled Firm Strategy: MASEM Procedure

Hypothesis 2 addresses FCFs’ strategic differentiation. To test it, we used meta-analytical structural equation modeling (MASEM; Bergh et al., 2016). The MASEM procedure combines the techniques of structural equation modeling with those of meta-analysis (Viswesvaran & Ones, 1995). To differentiate between strong and weak FBL societies, we categorized all 83 countries in our sample using a k-means cluster analysis (Punj & Stewart, 1983) on the five sub-dimensions of the FBLI measure. The k-means cluster routine assigned all countries up to an FBLI score of 0.50 to the “weak FBL” group (including the borderline case of Hong Kong), and all countries exceeding that threshold to the “strong FBL” group.

MASEM involves a two-stage procedure. In the first stage, we built a meta-analytic correlation matrix using Hedges and Olkin-type meta-analysis (HOMA; Hedges & Olkin, 1985). We used Pearson product-moment correlations (r) as effect sizes to compute the meta-analytic mean associations between all variables intended for inclusion in the MASEM analysis. Each effect size is weighted by its inverse variance weight w, the inverse of the squared standard error (Hedges & Olkin, 1985). These weights are used to compute the standard error of the mean effect size and confidence interval. To account for potential heterogeneity in the distribution of effect sizes, we obtained mean effect sizes using random-effects HOMA. We retrieved r from the correlation tables in the primary studies. In the second stage, we treated the meta-analytic correlation matrix as the observed correlation matrix and subjected it to regular maximum likelihood structural equation modeling routines (Duran et al., 2016). To deal with sample size differences across the correlation coefficients included in the matrix, the analysis is based on the harmonic mean sample size (Landis, 2013). The harmonic mean is less sensitive to outliers than the arithmetic mean, which yields more conservative t-values for the model parameters (Geyskens, Steenkamp, & Kumar, 2006).

Testing Hypothesis 2 requires us to assess, in both strong and weak-FBL contexts, the effect of family control on strategic choices commonly used in prior literature (van Essen et al., 2015): diversification, leverage, innovation, internationalization, and risk. In the MASEM analysis, we also controlled for firm size (measured as total assets, sales, or employees) since it is known to affect strategic decisions (Tihanyi et al., 2019). We tested this system of simultaneous equations on the two groups of countries – with strong and weak FBL.

Family-Controlled Firm Performance: MARA Procedure

Hypothesis 3 is concerned with FCF performance. To test it, we used meta-analytic regression analysis (MARA; Lipsey & Wilson, 2001). MARA is a weighted least squares technique that models previously unexplained variance in the effect size distribution (Lipsey & Wilson, 2001). One of the advantages of MARA is that it allows for the inclusion of theoretically derived country-level variables, such as our FBLI measure, that were not included in the primary studies (Arregle et al., 2017). MARA models treat the inclusion of these variables as moderators of the focal relationship (Lipsey & Wilson, 2001). Weighted regression was used to account for differences in precision across effect sizes. The statistically preferable weighting variable is, once again, w (Hedges & Olkin, 1985).

MARA uses the associational strength of the relationship between two variables as its dependent variable (Lipsey & Wilson, 2001). In our study, the dependent variable is the associational strength of the relationship between FCF and firm performance. Firm performance is measured as a latent multidimensional construct (Miller, Washburn, & Glick, 2013) that includes both accounting (ROA, ROE, ROS, earnings per shares, profit margin, and sales growth) and market-based (percentage of stock returns, Tobin’s Q, and market-to-book value) performance indicators. For the MARA procedure, we used partial correlation coefficients (rxy.z) as effect sizes, since the statistical properties of rxy.z allowed us to account for endogeneity, non-linear effects, and omitted variable bias (Stanley & Doucouliagos, 2012). We retrieved rxy.z from the t-statistics and degrees of freedom found in the primary studies (Greene, 2003).

Additionally, we included several control variables in the MARA models. At the country level, we controlled for formal institutions, stock market capitalization to GDP, the prevalence of FCFs in the country (Li et al., 2007), and ln GDP per capita. We also included dummies for the definitions of family firms (ownership and/or management definition as reference category). We also assessed whether the FCF was controlled by family members belonging to the first generation, later generations, or mixed generations (unknown family generation). We also control for the chosen operationalization of firm performance: market-based or accounting-based (reference group), adjusted for industry performance or not (reference group), and logarithmically transformed or not (reference group). To control for methodological artifacts, we tested for the “file drawer problem” (Rosenthal, 1979) by including a dummy variable indicating whether a study was published or not (reference group). To account for publication outlet status effects, we included each publication’s 5-year ISI impact factor. To allow for the possibility that the focal relationship is changing over time, we controlled for the median year of sample window. We also included dummies indicating whether effect sizes were based on a panel or cross-sectional (reference group) design and whether they derived from a study controlling for endogeneity of family control on firm performance or not (reference group). Finally, we included dummies to assess whether primary studies controlled for industry effects or not (reference group) and year effects or not (reference group).

We tested the robustness of the results against several model specification artifacts. We controlled for the number of variables included in the regression. We incorporated a dummy variable indicating whether independent variables were included as interaction effects or not (reference group) in the regression models of the primary studies. We accounted for omitted variables by including dummies indicating whether the following variables were included in the primary studies or not (reference group): firm advertisement expenditures, firm capital expenditures, firm diversification, firm dividends, dual listing, firm age, firm growth, firm risk, firm size, firm free cash flow, percentage of firm internationalization, firm debt, market risk, prior firm performance, percentage of R&D expenditure, affiliated with a business group, percentage of outside directors, board size, CEO duality, corporate ownership, foreign ownership, government ownership, inside ownership, institutional ownership, percentage ownership of the largest owner, second blockholder, and dual-class shares. Each of these variables was included as a control in at least five percent of all primary studies. Appendix D describes the variables included in the models.

RESULTS

Censored Regression Results: Family Business Legitimacy and Family-Controlled Firm Prevalence

Hypothesis 1 predicted a positive association between countries’ FBL and FCF prevalence. To compute FCF prevalence scores, we excluded studies with matched-samples designs, mixed-country samples designs, and studies that did not provide descriptive statistics on FCF prevalence. From the remaining 254 primary studies in our sample, we determined country-level FCF prevalence scores by taking the sample size-weighted arithmetic mean fraction of FCFs in the primary samples. FCF prevalence scores range from 0.08 (Ireland) to 0.73 (Greece), with a mean of 0.46 (see Appendix E).

Table 2 reports the Tobit regression results divided by FCF definitions. Models 1a and 1b report aggregated results, models 2a and 2b show ownership definition results, and models 3a and 3b show management definition results. Model 4 contains the aggregated results, broken down by FBLI dimensions. The results in Table 2 generally support Hypothesis 1: after controlling for formal institutions and stock market capitalization to GDP, the FBLI measure is a significant predictor of FCF prevalence in Models 1b and 2b. The results in Model 4 show that the network-based relations dimension of the FBLI measure is the strongest driver of FCF prevalence, indicating that family firms are especially ubiquitous in contexts in which FCF leaders are effective in building community-level social capital (Lester & Cannella, 2006).

Table 2 Results for prevalence of family-controlled firms (Tobit regressions)
Full size table

MASEM Results: Family Business Legitimacy and Family-Controlled Firm Strategy

To test Hypothesis 2, we created separate meta-analytic correlation matrices for both the weak- and strong-FBL subsamples using the r-based effect size distribution (see Appendixes F1 and F2). To maximize the statistical power of the MASEM tests (Hedges & Pigott, 2001) and to capture the greatest possible amount of FCF heterogeneity (Chrisman & Patel, 2012), we used the most inclusive definition of the FCF available: ‘management and/or ownership’ and ‘first or subsequent generation.’ To detect any FCF definition-based moderating effects (Miller et al., 2007; Villalonga & Amit, 2006), we control for FCF definition in subsequent multivariate analyses, such as those presented in Table 5.

Tables 3 and 4 contain the MASEM results. The MASEM models fit the data well, both the one estimated for weak FBL contexts (χ2 = 515.23; GFI = 0.96; RMSR = 0.069) and that for strong FBL contexts (χ2 = 119.11; GFI = 0.99; RMSR = 0.029). The results corroborate Hypothesis 2. In weak-FBL environments, FCFs are strategically isomorphic to non-FCFs (see Table 3; cf. Miller et al., 2012). However, in strong-FBL contexts, several significant deviations are observed (see Table 4) – FCFs are less leveraged than non-FCFs, spend less on R&D, and are generally more likely to avoid risk. FCFs thus only enact more conservative patterns of decision-making in contexts where they have the leeway to pursue their unique interests, thereby deviating from the strategies used by non-FCFs (Sherer & Lee, 2002). Conservatism is, therefore, best seen as a contingent characteristic of FCFs rather than a universal trait (Deephouse, 1999; Salomon & Wu, 2012).

Table 3 MASEM results for countries classified as weak family business legitimacy
Full size table
Table 4 MASEM results for countries classified as strong family business legitimacy
Full size table

MARA Results: Family Business Legitimacy and Family-Controlled Firm Performance

The MARA models testing Hypothesis 3 are presented in Table 5 (based on the rxy.z-based distribution). Model 1 contains study-level control variables and model specification artifacts (the content of the z-vector of control variables for each rxy.z-based effect size). Model 2 adds four country-level controls: formal institutions, stock market capitalization, the prevalence of FCFs in a country, and GDP per capita. Model 3 is the full model, which includes the FBLI measure and Model 4 is a full model including the five formative components of the FBLI measure as predictors.

Table 5 Results of mixed-effects WLS regressions
Full size table

The results of Model 3 support Hypothesis 3: the FBLI measure has a positive and significant moderating effect on the family control–firm performance relationship (β = 0.04; p < 0.05). In other words, FCF performance is stronger in strong FBL contexts than in weak FBL contexts.

Model 1 results suggest limited effects of definitional artifacts on the focal relationship. No significant moderating differences could be detected between the ownership- and management-based definitions. However, FCFs controlled by later generations showed weaker performance than those controlled by the founding generation, possibly because members of successor generations did not face similarly rigorous levels of selection and competition before assuming leadership positions as those experienced by members of the founding generation. Also, FCFs do slightly better in terms of market valuation as compared to accounting-based profitability. The influence of methodological moderators is similarly modest. We included many unpublished studies (36.16% of all retrieved studies), allowing for a reliable test, but no publication bias was noted. Also, the mean effect for studies with endogeneity corrections is similar to that of uncorrected studies, indicating that firm performance does not endogenously determine family control. We found a significant positive effect for journals with higher impact factors, indicating that studies with weaker results may find their way into print but perhaps end up in less prestigious journals. We similarly found a significant positive effect for the median year of the sampling window, suggesting that more recent studies report stronger effects. Such findings are usually attributed to methodological advancements over time (Mutlu, van Essen, Peng, Saleh, & Duran, 2018). Effects also tend to be weaker in studies with panel designs and stronger in studies controlling for year effects.

Importantly, Model 2 results for formal institutions do not support the thesis commonly found in law and finance that family firms do comparatively better in environments with weak formal institutions. Furthermore, neither the GDP-adjusted capitalization of the stock market nor the prevalence of FCFs in the country has a significant moderating effect on the family control–firm performance relationship. In contrast, the level of economic development of countries, as measured by ln GDP per capita, has a positive moderating effect.

Model 4 results show that three sub-dimensions of the FBLI measure positively moderate the focal relationship: continuity orientation (Miller et al., 2003), where FCFs are at liberty to enact patterns of strategic decision-making that honor their traditional ways of doing business; network-based relationship contexts (Lester & Cannella, 2006), where family members can pursue strategies that convert their community-level social capital into financial capital; and in-group solidarity societies (Fukuyama, 2002), where kinship ties enable family members to engage in value-adding exchange strategies that would be impossible to initiate with out-group members. We found counter-hypothesized effects for the intergenerational survival-oriented values and patriarchal domination dimensions. It might be that in societies high on these dimensions, FCF profitability is traded off against FCF longevity and the preservation of paternalistic dominance structures. As a robustness check, we reran all our analyses on an effect size distribution containing only r-based effects and on the combined distribution of r and rxy.z-based effects. The results remained identical, and are available upon request.

Robustness Checks and Post-hoc Analyses

We treat firm performance as a latent multidimensional construct (Miller et al., 2013). Since we know which performance indicator was used for each individual effect size, we can test the robustness of our findings against the various operationalizations chosen by different scholars. We included a separate dummy variable for each performance indicator in our MARA tests (Model 3, Table 5). We find that the focal relationship is stronger where performance is measured by earnings per share, profit margin, and market-to-book indicators, and weaker when it is measured as ROS. Several indicators yielded insignificant coefficients: ROA, ROE, ROI, sales growth, stock performance, and Tobin’s Q. We also tested whether Hypothesis 3 results remain stable when firm performance is measured only as accounting- or market-based performance. Hypothesis 3 is supported for accounting-based performance but rejected for market-based performance. This suggests that the constitutive legitimacy of FCFs in strong FBL contexts helps them gain market-power advantages (Morck, Wolfenzon, & Yeung, 2005), resulting in either in higher sales growth or in higher profitability through lower salary (Lubatkin, Ling, & Schulze, 2007) and innovation expenses (Duran et al., 2016). Formal institutional strength has no effect on the family control – accounting-based performance relationship but negatively moderates the family control–market-based performance association. This suggests that FCF reputation and formal institutions are governance substitutes: external shareholders rely on FCF reputation as a safeguard in contexts with underdeveloped formal institutions, but this safeguard becomes less important as formal institutions develop (Peng & Jiang, 2010).

We used several procedures to address potential endogeneity issues. First, we reran our analyses for Hypothesis 1 by disaggregating the time-invariant country-level prevalence scores (N = 45) to sample-level prevalence scores (N = 327), thus engaging in a pooled cross-sectional analysis. Because our samples cover a period ranging from 1953 to 2011, we could model the longitudinal country-level variance in terms of FCF prevalence and regress this variance on lagged time-varying predictor variables. The results of this disaggregated and lagged analysis are strong and comparable to those reported in Table 2. Similarly, to mitigate reverse causality issues for the other hypotheses, we only included effect sizes based on lagged or same-year measures of FCFs on both strategy variables (Hypothesis 2) and firm performance measurements (Hypothesis 3) (Oxelheim & Randøy, 2005; Post & Byron, 2015). Second, to check whether Hypothesis 3 results were prone to endogeneity bias, we conducted a separate HOMA on the subsample of endogeneity-corrected rxy.z-based effects (Jeong & Harrison, 2017). Appendix E shows that the mean effect size difference between this subsample and the full sample of partial linear correlations is trivial, thus suggesting that endogeneity issues are not a concern for Hypothesis 3 results. We confirmed this by including an endogeneity check dummy in Models 1 to 4 of Table 5. We find statistically insignificant coefficients for this control variable across the models. Finally, since omitted bias is likely to be present in the models we used to test Hypothesis 1 and 3 (Cao et al., 2018), we included country-level control variables, such as formal institutions and stock market capitalization, that might also affect FCF prevalence and the family control–firm performance relationship (Peng & Jiang, 2010). We also controlled for the number of variables in the regression and for the presence of 27 firm-level variables in the z-vector of each rxy.z-based effect to account for omitted variables in our test of Hypothesis 3 (Duran, van Essen, Heugens, Kostova, & Peng, 2019).

Since the institution-based view emphasizes that both formal and informal institutions affect firm behavior and outcomes (Peng, Wang, & Jiang, 2008), a salient question is whether the institutional mechanisms shaping FCF prevalence and performance are complements or substitutes (Helmke & Levitsky, 2004). To address it, we included an interaction term between formal institutions and FBLI in the models used to test Hypotheses 1 and 3. Results show a nonsignificant interaction effect on FCF prevalence. We did find a positive and significant effect on the family control–firm performance relationship, however. This suggests that formal and informal institutions complement each other, meaning that FCFs are best equipped to leverage their constitutive legitimacy in high FBL countries that exhibit favorable regulatory conditions for investors. We also explored the interaction effect between formal institutions and the FBLI measure on FCF strategic differentiation (Hypothesis 2). Specifically, we classified countries into four groups: (1) weaker formal institutions and lower FBLI scores, (2) stronger formal institutions and lower FBLI scores, (3) weaker formal institutions and higher FBLI scores, and (4) stronger formal institutions and higher FBLI scores. We then ran a separate MASEM for each group. Results show that FCFs located in countries with stronger formal institutions and higher FBLI scores exhibit strategic differentiation, as evidenced by lower debt, lower innovation, and lower risk relative to non-FCFs. No strategic differences were found in the other three groups, suggesting that FCFs are primarily prone to strategic differentiation in high FBL environments with strong formal institutional infrastructure (e.g., well-protected minority shareholders’ interests).

Finally, we explored the connection between FCF prevalence, strategic differentiation, and financial performance in high versus low FBL environments by splitting our sample into four groups of countries: (1) low FBLI and low FCF prevalence, (2) low FBLI and high FCF prevalence, (3) high FBLI and low FCF prevalence, and (4) high FBLI and high FCF prevalence. We then ran MASEM models for each group. In low FBL/low FCF prevalence countries, FCFs show strategic conformity, which translates into superior performance. Thus, in environments where FCFs are less legitimate and are not the dominant ownership type, institutional (Miller et al., 2013) and agency (Anderson & Reeb, 2003) theories accurately predict the strategic behavior and performance of FCFs. In line with Hypothesis 2, we find that FCFs show strategic differentiation in countries where family control is more legitimate, independently of whether FCFs are prevalent (group 4) or not (group 3). In line with socioemotional wealth argument, however, this strategic differentiation impedes FCFs from achieving superior performance (Martin & Gómez-Mejía, 2016). Interestingly, we find that FCFs internationalize more than non-FCFs in low FBL/high FCF prevalence environments (group 2) suggesting that social pressures to act in stewardly ways lead FCFs to pursue international opportunities (Zahra, 2003), probably because their dominant position in the domestic arena reduces home-country barriers to internationalization. All results are available from the authors upon request.

DISCUSSION

Theoretical and Empirical Contributions

Our study provides needed clarity, depth, and nuance to the FCF literature, which on the one hand has formed a steady narrative around family firms, while on the other hand it has challenged some of this narrative by documenting notable exceptions and inconsistencies, especially in comparative research. The narrative focuses on the distinctiveness of family firms – they pursue strategies that are likely to benefit the family across generations (Bennedsen et al., 2010; Chua et al., 1999), for example by underinvesting in risky activities like exploratory R&D (Duran et al., 2016); they lean toward conservative stakeholder-oriented strategies that protect the family socioemotional wealth (Cennamo, Berrone, Cruz, & Gómez-Mejía, 2012); and as a result of such prudent and stewardly approaches, they often outperform non-FCFs led by more aggressive, self-serving, and short-term oriented professional managers (Miller, Lee, Chang, & Le Breton-Miller, 2009). A number of studies, however, have documented cross-country variability of FCFs’ strategic behaviors and performance, which challenges these “truisms.”

Our study aimed to explain and reconcile these inconsistences. We combined insights from the contextual embeddedness perspective with the power of the meta-analytical methodology to provide additional depth and nuance to the literature. Our main thesis, strongly supported by the empirical evidence we mustered, is that the cross-country variability in FCF strategies and performance can partly be explained by the contextual effects of national informal institutions, particularly by what we called family business legitimacy (FBL), which captures the societal perceptions about family business as an organizational form. We find that a country’s FBL affects the prevalence, strategic differentiation, and performance advantages of family firms. In essence, we show that strong FBL is a boundary condition for many of the findings in the family business literature; thus, they may not be valid in countries with less appreciation for the institution of family business. We believe that several of our findings offer interesting empirical and theoretical insights.

Regarding the prevalence of FCFs, we show wide variability – in some countries, they constitute the majority of listed firms, but in others they are only a minority (see Appendix E). To explain this variance, previous studies have offered a “substitution” hypothesis arguing that family ownership and control can compensate for weak formal institutions in a country (e.g., Djankov et al., 2007, 2008; La Porta et al., 1998). We find no support for this argument in our meta-analysis: a composite variable capturing various formal institutional indicators did not yield any significant effects on FCF prevalence. Instead, we find a significant effect for countries’ FBL, representing their informal institutions. Where a country is supportive of family firms, entrepreneurs are likely to invest in FCFs, and FCFs are likely to have higher funding and support and lower failure rates. In contrast, in countries less favorable to family business, entrepreneurs tend to invest in non-FCF organizational forms, leading to low prevalence of FCFs. These findings underscore the importance of informal institutions in the context of family business.

Our study also sheds light on the question of strategic differentiation of family businesses. Several scholars have examined FCFs’ susceptibility to isomorphic forces (cf. Aguilera et al., 2018; Miller et al., 2012). In a study of U.S. FCFs, Miller et al. (2012) found that these companies are rather sensitive to such pressures, displaying high levels of strategic isomorphism with their non-FCF counterparts. The authors explain that family firms are “subject to unusually powerful motivations to conform, in part because of their pursuit of socioemotional wealth objectives” (p. 189) and may conform to strategies that are more suitable for firms with dispersed ownership, foregoing family-benefiting strategies that they would otherwise prefer. We confirm Miller et al.’s findings, but we offer an alternative explanation. According to our analysis, U.S. is part of the “weak FBL” cluster of countries (see Table 1), in which family firms do not show significant differentiation from non-family firms (see Table 3). It is the opposite in strong FBL countries. There, FCFs seem to be granted a “license to be different” as evidenced by the significant differentiation on several strategic dimensions (see Table 4). Therefore, the FBL construct is a powerful contingency that captures important informal institutional conditions determining whether FCFs will engage in strategic conformity or strategic differentiation.

Additionally, our post hoc analyses show that FCFs follow strategic differentiation only in countries with high scores on both FBL and formal institutions. This is consistent with the argument that organizational discretion requires complementary institutions to act in concert (Crossland & Hambrick, 2011). We also speak to strategic balance theory (Deephouse, 1999; Salomon & Wu, 2012), which suggests that firms need to do both – seek strategic isomorphism to gain legitimacy and strategic differentiation to compete. We show how institutional contingencies like FBL create an additional set of forces around this tension – allowing or constraining firms from exerting differentiation and/or isomorphism (cf. Salomon & Wu, 2012). The strength of a country’s FBL can, therefore, explain the extent to which FCF strategic similarity and differentiation can coexist.

Finally, our work provides additional insights into the performance implications of family ownership and control. Several high-impact studies have found a positive performance differential for FCFs compared to non-FCFs (cf. Amit & Villalonga, 2014), which has been attributed to the motivations and management styles of FCF leaders prioritizing the long-term financial and socioemotional wellbeing of the family (Miller & Le Breton-Miller, 2009). Our study further qualifies this finding by showing that the performance advantages of FCFs are, in part, context-dependent. The ability of FCF leaders to outperform is moderated by the FBL level of the country – they act as stewards when they can and act as agents when they must. We see an opportunity to build on this insight with future behavioral agency research (Wiseman & Gómez-Mejía, 1998) examining the impact of FBL informal institutions on the agentic versus stewardly decision-making patterns in FCFs. Lastly, post hoc analyses show a positive interactive effect of formal and informal institutions on FCF performance. Again, this is a departure from the idea that FCFs’ performance advantages are driven by the exploitation of institutional weaknesses and that informal institutions mostly emerge as substitutes or ‘compensatory structures’ for weak formal institutions (cf. Peng & Jiang, 2010; Peng et al., 2018). In contrast, our study suggests that formal and informal institutions vary independently, but can reinforce each other’s positive effects on FCF performance in institutional configurations in which both are strongly developed.

Practical Implications

Our work has a number of implications for FCF leaders. Understanding the ways in which informal institutions affect their business can inform their decisions about growth opportunities, the possibilities and limits for strategic differentiation, and the realistic performance advantages that they might expect to realize. Strong FBL countries allow more leeway, and operating there likely leads to better performance, but the lack of constitutive legitimacy of family business in weak FBL countries could be detrimental. There, FCFs are likely to be a minority organizational form and face strong conformity pressures that might wipe out their strategic advantages relative to non-FCF firms and disrupt the preservation of socioemotional wealth. In such environments, FCFs might struggle to realize adequate financial performance and eventually have to face existential choices. One option is to display strategic conformity and downplay FCF-specific behaviors. This might help in warding off stakeholder skepticism (Luo et al., 2019), but the cost could be a perennial erosion of socioemotional wealth. A more drastic option is to abandon the FCF organizational ownership structure, for example, by exiting the firm as owners or by merging with a non-FCF. While this may be more disruptive to socioemotional wealth in the short run, it could be a better strategy in weak FBL contexts. Yet another alternative is to explore international opportunities by locating in strong FBL countries; although it is not clear whether the constitutive legitimacy of family business in a given country will also extend to foreign FCFs.

Policymakers could also be informed by our study. Regardless of the favorability of FBL, family businesses create much value for a country’s economy and society, contributing to its employment, social stability, and economic development goals. It is, therefore, in their interest to promote family business to gradually improve the social perceptions towards it. However, this might not be politically expedient when their constituents harbor unfavorable attitudes towards it. They could consider a range of approaches to this end, for example, by providing factual information that underscores the benefits of this organizational form, and by investing in compensatory institutional arrangements with the potential to support FCFs in lieu of the lacking constitutive legitimacy. Introducing regulations and laws to curb some notoriously bad practices of family businesses, like tunneling, nepotism, and patriarchal domination might also be effective in improving social perceptions about FCFs. Policymakers could also strengthen respective formal institutions to compensate for FCF weaknesses, such as labor market institutions to address FCFs relatively weak reputation as employers or financial market institutions to facilitate FCFs access to external finance (Duran et al., 2019). These dedicated investments and initiatives could, to some extent, help level the playing field between FCFs and non-FCFs, even in weak-FBL settings.

Limitations and Future Research

We suggest some directions for future research. The first implication relates to sample selection practices. The FCF literature is less U.S.-centric than most other management literatures (which is interesting in itself and might reflect the fact that FCFs have a weaker informal institutional embeddedness in the U.S. in comparison to many other countries). Still, this greater variety of sampling contexts comes with its own difficulties. A meta-analytic hierarchical linear modeling exercise on the 484 studies included in our sample revealed an intra-class correlation coefficient of 0.42, which implies that 42 percent of the variability in FCF outcomes is determined by country-level factors. We, therefore, urge researchers who are designing international or comparative FCF studies to first consult the FBLI measure and the country scores listed in Table 1 before selecting their sample and making theoretical predictions. In particular, they should be aware that in weak FBL contexts, the baseline hypotheses should be that FCFs are outnumbered, strategically inconspicuous, and ordinary performers. In contrast, in strong FBL contexts, they can expect FCFs to be ubiquitous, strategically differentiated, and financial outperformers. We are hopeful that with the introduction of the FBL concept and the FBLI measure as a relevant contingency factor, we have contributed to improving the accuracy and contextual validity of future studies aiming to explain FCF prevalence, strategy, and performance.

Second, in the comparative management literature, there are persistent calls for more management-specific measures of institutional development (Gedajlovic et al., 2012; Peng et al., 2008). Such measures would complement existing metrics of institutional development like the rule of law index (Kaufmann, Kraay, & Mastruzzi, 2005) and the ease of doing business index (WorldBank), which are too generic and primarily intended to explain macro-economic growth. We note that several indexes have been developed that focus on the formal institutions necessary for firms’ effective functioning, such as creditor protection laws (La Porta et al., 1998) and shareholder protection laws (Djankov et al., 2008). We took a different approach and developed an institutional index that captures the informal institutions conducive to a specific organizational form – FCFs. Organizational form-specific metrics are an attractive alternative to market-focused indexes because they are closer to the phenomenon under study and are better able to explain the contextual contingencies operating on specific firms’ behavior and outcomes. We welcome further efforts to develop organizational form-specific indexes, focusing on organizational forms like business groups, professional service firms, or state-owned enterprises. We would also welcome novel empirical tests of the FBLI measure aimed at refining its explanatory potential. Since several of our analyses yielded insignificant or negative results for the patriarchal domination dimension (see Tables 2 and 5, respectively), more work is needed to determine whether four or five groups of indicators would provide the best possible specification of the index.

This study is also subject to some limitations. First, our sample is limited to publicly listed FCFs. The upside of this sampling strategy is that it makes our sample more comparable across national contexts, because all included firms meet certain thresholds in terms of size, financial structure, and organizational governance. At the same time, there is substantial variance across contexts with respect to the size and prominence of FCFs. These differences primarily pertain to the position of private family firms in the local economy, as the variance in this subset of FCFs is decidedly larger than that amongst public FCFs. While it is conducive to our hypothesis tests that the between-firm heterogeneity in our sample is lower than the heterogeneity in the universe of all private and public FCFs, an interesting question for future research is how FBL impacts the prevalence, strategy, and performance of private FCFs.

Second, in line with the institutional theory perspective we employed (DiMaggio & Powell, 1983; Meyer & Rowan, 1977), we argued that the informal institutional environment (i.e., FBL) would determine whether our focal organizational form (i.e., FCF) would enjoy societal support, which we portrayed as a precondition for it to thrive (Li et al., 2007; Luo et al., 2019). Through a number of additional analyses, we have sought to establish the directionality of this relationship, and have indeed found supportive evidence for our central premise that FBL drives FCF prevalence, ruling out to some extent the opposite conjecture that FBL is driven by FCF density-induced taken-for-grantedness. Given the limitations of using aggregated secondary data, however, we could not fully establish the direction of causality by correcting for endogeneity by using an instrumental variables design or by capitalizing on an exogenous shock. An important opportunity for future research, therefore, consists of using the FBLI we developed on primary data, using strong instruments to further disentangle the relationship between FBL and FCF prevalence.

Notes

  1. 1

    The appendices are available as online supplements.

  2. 2

    Specifically, the FBLI is negatively correlated with the variables that compose the formal institutions measure, such as non-French origins of the legal system (r = − 0.21), the anti-self-dealing index (r = − 0.10), creditor rights (r = − 0.16), minority shareholder rights (r = − 0.18), and judicial efficiency (r = − 0.58).