Family and Lone Founder Ownership and Strategic Behaviour: Social Context, Identity, and Institutional Logics

Authors


abstract

There is controversy in the literature about the effects of ownership on strategy and performance. Some scholars have taken agency explanations as definitive, arguing that closely held firms outperform. Empirical studies, however, show conflicting findings for firms with concentrated ownership: lone founder firms outperform, family firms do not. Such conflicts may be due to the failure of agency theory to distinguish between the social contexts of these different types of owners. We argue that explanations of performance must take into account not simply ownership, but who are the owners or executives and how their social contexts may influence their strategic priorities. Family owners and CEOs, influenced by family stakeholders in the business, are argued to assume the role identities and logics of family nurturers and thus strategies of conservation. By contrast, lone founders, influenced by a wider set of market-oriented stakeholders, are argued to embrace the identities and logics of entrepreneurs and strategies of growth. Family founders and founder-executives are held to blend both orientations. These notions are supported in a study of Fortune 1000 companies.

INTRODUCTION

There is lively debate in the literature about the effects of ownership on strategy and performance (Durand and Vargas, 2003; Greenwood et al., 2007). Some scholars have taken agency explanations as definitive, arguing that closely held firms outperform because they can curb opportunism by agents (Jensen and Meckling, 1976). Others have taken exception to that view, suggesting that ownership concentration has negative agency effects on strategy and performance as major owners use their power to exploit the business (Morck et al., 1988, 2005).

These disputes in the governance literature exist even when studying concentrated ownership of the personal variety. Some researchers find that family firms outperform (e.g. Anderson and Reeb, 2003; Villalonga and Amit, 2006), others that they do not (e.g. Bennedsen et al., 2007; Maury, 2006). Miller et al. (2007), in distinguishing family firms from lone founders firms, showed these differences to be attributable to the identity of the owners. Thus predictions of agency theory that rely on the positive effects of ownership concentration do not apply across the board: the performance of family firms and lone founder firms, both of which share concentrated ownership, is quite different. The social context of owners, not simply their share ownership, seems to matter. We shall argue therefore that it is useful to go beyond agency theory – which is too narrowly focused on economic self interest – to examine the social contexts of these owners and the interaction constituencies, role identities, and behavioural priorities to which these divergent contexts give rise (Gomez-Mejia et al., 2001; Greenwood et al., 2007). These latter characteristics are encapsulated by a literature on ‘institutional logics’, specifically on ‘familial’ and ‘market’ logics that we argue pertain to family and lone founder actors, respectively (Friedland and Alford, 1991; Thornton et al., 2005). We examine the concrete business strategies suggested by these contexts in a sample of Fortune 1000 companies.

Governance Disputes and the Limits of Agency Theory

According to agency theory, ownership structure can affect managerial behaviour and firm performance. Businesses in which ownership is concentrated are said to outperform because their owners have the power and incentive to better monitor their managerial agents and thereby avoid the information asymmetries that raise agency costs (Jensen and Meckling, 1976; Morck et al., 1988). Major owners also are said to exploit smaller ones by entrenching themselves and appropriating firm assets (Morck et al., 2005).

There have, however, been conflicts in the literature on governance that suggest the inadequacy of an agency perspective to explain the consequences of ownership concentration. These conflicts become clear when we examine the last several decades of research. Some research on large public companies showed that family firms outperform. For example, Anderson and Reeb (2003), Kang (1999), Leach and Leahy (1991), McConaughy et al. (1998), and Villalonga and Amit (2006) found that family firms outperform in market valuations, and often also in return on assets, especially during the first generation. By contrast, other studies found that family firms do not outperform, and may even underperform, despite any presumed agency advantage (cf. Bennedsen et al., 2007; Cronqvist and Nilsson, 2003; Holderness and Sheehan, 1988; Maury, 2006; Pérez-González, 2006). In attempting to reconcile these findings, Miller et al. (2007) found in their analysis of US public firms that whereas lone founder firms, which previous research had classified as family firms, did enjoy superior market valuations, family firms did not – even in the founder generation. The crucial factor determining performance was whether or not an individual lone founder was present or a set of owners or managers acting with other family members as major shareholders or officers in the company. In both types of organizations, ownership was concentrated, yet performance differed – an outcome neither anticipated nor explained by agency theorists.[1] A major difference here is the social context of ownership as reflected by key constituencies – specifically, the presence of multiple influential family actors in one type of firm, and a more typical set of arms length investors and stakeholders in the other. In family firms, family actors are primary constituencies of one another. Their close affective ties may evoke in them familial attitudes and agendas even in a business context. By contrast, lone founders' constituencies are often more emotionally detached, diverse, and financially motivated. They will encourage more of a commercial agenda.

There are many ways of characterizing social context and ours is driven by contrasts between the settings of lone founders and family actors. As the social constituencies of these parties differ, so will their consequent role identities, sources of legitimacy, and thus strategic priorities. The institutional literature encapsulates these categories in the form of two common ‘institutional logics’ adopted by key owner-actors: the familial logic of family owners or owner executives and the market or ‘entrepreneurial’ logic of lone founders (Friedland and Alford, 1991; Thornton et al., 2005).

Thornton and Ocasio (1999, p. 804) define institutional logics as ‘the formal and informal rules of action, interaction and interpretation that guide and constrain decision makers. Thus organizational strategies and practices are legitimized by, and also manifestations of, such logics’ (Greenwood et al., 2010; Reay and Hinings, 2005). According to Thornton (2004, p. 2) logics ‘emerge from the institutional orders of the inter-institutional system’, namely, family, religion, democracy, and market capitalism. Logics, because they confer legitimacy, can mould behaviour. They have the most influence when individuals ‘identify with the collective identities of an institutionalized group’ such as a family or profession (Thornton and Ocasio, 2008, p. 119). Although identity is a function of social interaction with salient constituencies (Stryker, 1980), it also may arise from a sense of commonality or an aspiration, with or without regular social contact with a reference group (Tajfel and Turner, 1979). Behaviour opposing legitimated logics may be punished socially.

The Social Context of Ownership: Interaction, Identity, and Logics

The social context of owners can influence their identities and logics in several ways. One is simply through the use of common cognitive frames (Burke, 1980). Owners, through frequent association with a stable group of people, come to share perspectives – they share scripts for understanding the world: scripts about ‘family loyalty’, for example (Phillips et al., 2004). Social context may also have a normative impact – as reciprocation and emotional closeness breed a sharing of values and a sense of responsibility to the group: thus family owners may behave altruistically towards their kin (Schulze et al., 2001; Stryker, 1980, 1987). Finally, are the more readily observable political dimensions of social context whereby identification with a group and dependence upon it bring pressures to follow a group agenda: a founder must serve key investors (Karra et al., 2006; Greenwood et al., 2010). These aspects of social context give rise to role identities consistent with group norms (Ashforth and Mael, 1989; Stryker, 1980, 1987). They also induce owners to assume societal logics closest to the aspirations of the group (Thornton, 2004). Such cognitive and normative influences are exerted in greatest measure by proximate groups such as a family, but also in part by more amorphous ones, such as an elite an actor identifies with – for example, the class of entrepreneurs (Tajfel and Turner, 1979).

We can relate social context, role identities, logics, and strategic behaviour more concretely by moving to specifics. Take first a major family business owner or owner-executive. That individual must interact with other family owners, directors, or managers in the business. Given the intimacy, stability, emotional impact, and broad scope of family ties, other family members typically will represent disproportionately influential constituencies who may bring to bear significant cognitive, normative, and political influences (Nisbet, 1970). Family members often demand from their firm steady financial support, security, careers, and ‘altruistic’ benefits to achieve family harmony (Schulze et al., 2001). In so doing they call forth among family officers familial identities and family nurturing roles (Stryker, 1980), and bestow legitimacy to those serving family wants (Bertrand and Schoar, 2006). These orientations reflect the familial logic of Friedland and Alford (1991) – one of nurturing, generativity, and loyalty to the family. They motivate family owners to manage the business to provide family members with stable incomes, long term security, and control of the firm (Morck et al., 2005; Schulze et al., 2001). Such priorities can shape business strategy by diverting firm resources to serve family needs and enforcing financial conservatism. They favour generous payouts and avoid risky projects or investments that might jeopardize family income or control. In short, familial contexts, role identities, and logics may elicit a ‘conservation’ strategy; one that may well limit performance.

The social context of the lone founder is quite different. There are no family members present as compelling constituents but rather a more diverse set of stakeholders – venture capitalists, partners, investors, employees, and customers – whose interests are largely economic. Via regular interactions, these parties exert their influence on the founder (Donaldson and Preston, 1995). There are pressures to reap returns for investors and partners, to expand opportunities for employees, to enhance service to clients, and to measure up to competitors (Hitt et al., 2002). In addressing these demands founders must embrace an entrepreneurial role – one that secures legitimacy by serving these economically-motivated constituencies (Loasby, 2007; Stryker, 1987). Lone founders also may identify symbolically with the role of business builder or creator, a role that develops in the course of various business initiatives and interactions with fellow entrepreneurs (Loasby, 2007; Tajfel and Turner, 1979). These identities, roles, and sources of legitimacy relate to what Boltanski and Thevenot (1991) and Thornton et al. (2005) variously term the ‘market logic’ or ‘capitalist logic’ that prizes growth in share price, wealth accumulation, keen competition, and committing investment capital. As such we call it an ‘entrepreneurial logic’. This logic endorses, even celebrates, a strategy of growth that invests in the business to grow and improve it and seize opportunities.

We will elaborate briefly on these orientations of lone founder and family owners and executives (summarized on Table I), and draw hypotheses about their firm's strategies and performance. Our methods and findings will then be presented and discussed.

Table I. Lone founder and family orientations in public enterprise
Orientations Lone founder firm Family firm
Type of major owner or top executiveLone founderMember of founding family
Salient social constituenciesBusiness stakeholders and other entrepreneursOther family owners, officers, and members
Role identityEntrepreneurFamily nurturer
Source of legitimacyGrowth, wealth accumulationFulfilment of family needs
LogicEntrepreneurialFamilial
Favoured strategyGrowthConservation
PerformanceSuperior shareholder returnsInferior shareholder returns

ORIENTATIONS OF MAJOR PERSONAL OWNERS AND EXECUTIVES

The Entrepreneurial Orientation: Lone Founders and their Firms

Context, identity, and logic.  Given a social context of constituencies such as venture capitalists, investors, partners, and managers, lone founders are expected to embrace the entrepreneurial role identity and its associated logic. As noted, the entrepreneurial identity may be reinforced by a founder's frequent interactions with these stakeholders as a business builder and deal maker, and as a participant in struggles and successes achieved over the years (Stryker, 1980). Many of these stakeholders will demand superior growth in return for their investments in an emerging enterprise (Knight, 1921). For example, investors and venture partners will want ample and growing returns, managers will want advancement. Thus a founder's accountability to these stakeholders will support an entrepreneurial mission. The legitimacy and authority, even self-esteem, of a lone founder may derive in part from these stakeholders and that mission.

Lone founders also may assume an entrepreneurial role identity because they see similarities between themselves and other firm founders (Tajfel and Turner, 1979). Indeed, that identity may be influenced by how other founders do their jobs, and even by the way they are portrayed in popular culture and the press as economic forces, business builders, innovators, risk takers, and virile competitors (Loasby, 2007; Miller, 1983). In short, the entrepreneurial logic follows from the assumption of an entrepreneurial identity. Its themes centre on commercial venturing and innovation, market opportunities and competition, and above all, capital gains and growth for the firm and its investors (Thornton et al., 2005).

Strategy and performance.  Given their social context, identities, and logics, we expect that firms dominated by lone founders will adopt a mission and strategy of growth. The classic notion of a growth strategy has been much discussed in the strategy and entrepreneurship literature. It is associated with innovation, long term investment, market expansion, and ever increasing wealth for investors (Ansoff, 1965). A growth strategy has two complementary components – long term investment and financing that investment (Ansoff, 1965; Hofer and Schendel, 1978). Investments to develop the business include those for R&D, promotion, and capital equipment (Davidsson, Delmar and Wiklund, 2002; Hitt et al., 2002; Miller, 1983). They are funded by restricting dividend payments (Hofer and Schendel, 1978), debt financing, or keeping cash in the business (Hitt et al., 2002).

The question of whether any founder or family influence is due simply to major ownership or to ownership and management remains an open one. Certainly, founder (or family) owners have the power to shape their organizations (Shleifer and Vishny, 1997). It is possible, however, that a founder or family member occupying the position of CEO also may have substantial impact (Miller et al., 2007). Thus we formulate our hypotheses using major ownership alone, as well as service as CEO.

Hypothesis 1: Firms in which a lone founder is the largest shareholder or the CEO will be associated with a growth strategy – one that pursues superior investments in R&D, promotion, and capital expenditures, embraces leverage, builds up cash reserves, and eschews dividends.

Growth strategies are said to be associated with higher levels of shareholder returns (Knight, 1921; Miller, 1983). An ample reflection of these results would be the cumulative total returns from share price accumulation (Hitt et al., 2002). More generous and more far-sighted investment in the business allows a firm to build competencies and seize opportunities. Moreover, a founder-owner whose self image and social approval are tied to the success of the firm is apt to invest capital into projects that will benefit the business and its investors (Kirzner, 1979). We are not arguing that profit margins necessarily will be higher for growing firms as these are not the primary concern. Moreover costs can be high during periods of growth. However, generous investments in innovation, market development, and capital equipment can provide the foundations for growth for years to come, and should lead to enhanced revenues and profits, and thus increasing share prices (Chandler, 1990).

Hypothesis 2: Firms in which a lone founder is the largest shareholder or the CEO will be associated with above average returns to shareholders.

The Familial Orientation: Family Owners and Executives and their Businesses

Context, identity, and logic.  The role identities, priorities and logics of family owners and owner-managers are expected to be shaped by their enduring and often intimate associations with other influential family members in the business. For these owners, focal constituencies may be family members – parties who due to their kinship ties may exert disproportionate influence (Nisbet, 1970). Thus interactions with family and regular exposure to their needs and demands may elicit a nurturing role identity (Burke and Reitzes, 1981; Stryker, 1980, 1987; Ward, 2004). Indeed, a family agenda may become salient almost daily in the form of members' expressed points of view, needs, values, emotional ties, and constraints. In a business, archetypal family desires and aspirations arise concerning family careers, security, and rewards from the firm, and preserving the company for later generations (Miller et al., 2010). That family agenda becomes influential through current and past social intercourse among family owners and executives; indeed the legitimacy and authority of the latter may derive in large measure from their ability to satisfy other family members (Bertrand and Schoar, 2006). A resulting family nurturing orientation reflects Friedland and Alford's (1991) familial logic in which family needs rival economic purpose. Generosity thus may be accorded to family members using firm resources, even when those individuals contribute little to the business (Lubatkin et al., 2007).

Strategy and performance.  Strategy can be shaped by this agenda of providing stable, secure income and even careers to family members, and preserving family control. Such priorities restrict the resources available to a firm, depriving it of funds to invest in the future (Landes, 2006). Also, in seeking to stabilize cash flows, they limit the capacity to assume risk. Thus a conservative investment and financial posture is taken – one that defines a ‘conservation strategy’ (Harrigan and Porter, 1983). First, is a reluctance to invest aggressively in speculative or longer term initiatives such as research and development, promotion and advertising, and extensive capital expenditures (Bertrand and Schoar, 2006). As these investments may take long to pay off, they would deprive a family of income, and risk family capital and control of the business. So would debt and large cash holdings – both of which are rendered unnecessary by the modest investment profile (Mishra and McConaughy, 1999). Instead of reinvesting profits, they can be distributed to the family (Dreux, 1990).

Hypothesis 3: Firms in which family members are the largest shareholders or serve as CEOs will be associated with a conservation (i.e. non-growth) strategy – one that minimizes investments in R&D, promotion, cash holdings and capital expenditures, avoids leverage, and awards generous dividends.

Conservation strategies limit competency demands on family managers, reduce risk, and place more ample resources at the disposal of the family. They keep family members feeling secure and avoid initiatives that might jeopardize family control of the firm by current and later generations (Dreux, 1990; Mishra and McConaughy, 1999). Spare liquid asset profiles also lessen the chances of hostile takeovers, while generous dividends may placate family and other shareholders. Unfortunately, sparse investment in the business and its future is apt to diminish total shareholder returns. A dearth of investment in innovation, market development, and capital equipment can hinder growth over the years, limit future revenues and profits, and thus restrict increases in share prices.

Hypothesis 4: Firms in which family members are the largest shareholders or serve as CEOs will be associated with below average shareholder returns.

There is controversy in the literature regarding Hypothesis 3. Some scholars contend that family owners, concerned with firm longevity and future generations, invest generously in the business (Gomez-Mejia et al., 2007; James, 2006; Miller and Le Breton-Miller, 2005). But these studies are unsystematic and based mostly on private firms. Thus whether family businesses favour conservation over growth is very much an open question. There is equal controversy surrounding Hypothesis 4. As noted, some researchers believe that family firms outperform their peers along numerous measures, especially in the founder generation (Anderson and Reeb, 2003; Villalonga and Amit, 2006). Other studies show that family firms do not out-perform and even may underperform (Bennedsen et al., 2007; Cronqvist and Nilsson, 2003; Pérez-González, 2006). Unfortunately, except for Miller et al. (2007), who looked only at market valuations, all prior empirical studies of public US family businesses of the founder generation confound family founder and lone founder businesses. Our arguments concerning familial orientations apply only to the former as there are multiple family members who own the firm and thus evoke familial agendas and sources of legitimacy. By contrast, where a founder has no family involved, the entrepreneurial logic is likely to apply.

Family Founders: Blended Orientations within Family Businesses

Entrepreneurial and familial orientations have been argued to pertain, respectively, to organizations that are lone founder and family controlled. But what about firms that are controlled by a family founder– a founder who has other members of his or her family present as a major co-founder, owner, or executive. Within this special type of family firm, the social contexts, identities, and logics we have described may blend. Indeed, Glynn and Lounsbury (2005), Reay and Hinings (2005), and Thornton et al. (2005) make clear that individuals and organizations may embrace multiple logics. This tendency may be especially common where these parties act within several social spheres, and thus incorporate compound identities (Chung and Luo, 2008; Fiss and Zajac, 2004; Greenwood et al., 2010). Although most lone founders are expected to embrace the entrepreneurial logic, family founders may be influenced by both the institutions of market-capitalism and family, and thus blend entrepreneurial and familial logics. These parties may identify with both entrepreneurial and family roles, viewing themselves as business builders and family nurturers. As such, their strategic conduct, and the performance of their firms, will fall in an intermediate position vis-à-vis other family and lone founder firms.

Firstly, family founders are entrepreneurs – they have established and grown their business. Therefore they take pride in that achievement and have demonstrated the values and capabilities that have enabled them to succeed in a context of market capitalism: that is apt to affect their assumption of an entrepreneurial identity and reference group (Tajfel and Turner, 1979). Moreover, they have built up extensive relationships, and reciprocal loyalties, with non-family members associated with the business – investors, bankers, employees, and clients, all of whom might well reinforce an entrepreneurial logic. Family founders' sources of legitimacy and authority too are apt to come in part from these parties. Thus a strategy of growth for the enterprise and its stakeholders may hold some appeal for them.

However, unlike the lone founder, the family founder does have ties to other family members in the business – ties that, given the intimate and often emotional associations between family members, may be especially important (Aldrich and Cliff, 2003). So the family too becomes a reference group. A family nurturing identity may loom large. Kinship involvement may also create demands to divert resources and accord perquisites to the family, provide career opportunities for family members, and otherwise satisfy family owners who are close to the founder but more concerned with family rewards than business growth (Gersick et al., 1997; Lubatkin et al., 2007). In short, the familial context, identity, and logic too will influence family founders via its symbols, sources of legitimacy and authority, and reinforcing relationships.

Therefore the family founder, as a ‘man in the middle’, may incorporate and balance between the two contexts, identities, and logics. As a result, strategies will lie somewhere between the poles of growth and conservation. In short, family founder firms will be less likely than lone founder firms, and more likely than other family firms, to embrace a growth strategy, and hence less likely to outperform or underperform, respectively. Their intermediate position may render them no different in their strategic orientations from other public companies.

Hypothesis 5: Firms in which a family founder is the largest owner or serves as CEO will be neither more nor less likely to adopt a growth strategy than other companies.

Hypothesis 6: Firms in which a family founder is the largest owner or serves as CEO will be neither more nor less likely to outperform in shareholder returns than other companies.

METHOD

Sample

Our sample consisted of the Fortune 1000 (500 industrials and 500 service firms). We analysed data on 898 companies due to our restricting the sample to firms with publicly accessible data for the years 1996 to 2000; of these we identified 263 family firms, 141 lone founder firms, and 492 ‘other’ firms. This breakdown is consistent with that found by other scholars of major publicly traded American firms (Anderson and Reeb, 2003; Miller et al., 2007; Villalonga and Amit, 2006). Sample sizes varied somewhat with the use of lagged variables and more restrictive ownership and management definitions of lone founder and family firms.

Variables and Sources of Data

Our variables were assessed at two levels and in two stages. First, we gathered data on individual officers and directors, 5% block-holders, and large institutional investors. Information on share ownership, vote control, family and lone founder executive positions, super voting shares, etc. was obtained from three or more sources for each company: Compact Disclosure, individual company proxies (which were the primary and definitive source of data), Hoover's, and company websites. Wherever the proxies lacked information on kinship relationships between board members and managers, or officers' relationships with the family founder, we approached companies ourselves. For two years, a team of five research assistants and three of the authors gathered data from the proxies. Two weeks of training were required for each research assistant as some families controlled their firms via trusts and banks, and family names could change due to marriage.

Following Anderson and Reeb (2003), Maury (2006), and Villalonga and Amit, (2006), we adopted as the focal family the one with the most votes. In summing family shareholdings, we included shares of co-trustees of family trusts who were directly employed by the family. However, whereas previous researchers (e.g. Anderson and Reeb, 2003; Maury, 2006; Villalonga and Amit, 2006) designate firms like Microsoft and Amazon as founder family businesses, we do not since there is no family involvement. Instead, because of their hypothesized different institutional logics, we classify them as lone founder (or unrelated-founder in the rarer instances of multiple founders) businesses. By contrast, we do count firms such as Comcast and Qualcomm as family firms because there are multiple members of their owning families serving as major owners and officers. Lone founders had no relatives involved in their firm.

All data on individuals were aggregated to the firm level, at which we could also collect information on strategy, governance, and market performance. Accounting data are drawn from Compustat, and market performance data were obtained from the Center for Research on Security Prices (CRSP). Our variables are listed and defined in Table II, along with their sources; descriptive statistics are presented in Table III.

Table II. Variable definitions
Variable Definition
Lone founder largest ownerA binary variable, 1 indicates that a lone founder is the largest shareholder in the firm. Source: Compact Disclosure; firm proxy.
Family largest ownerA binary variable, 1 indicates that a family is the largest shareholder in the firm. Source: Compact Disclosure; firm proxy.
Family founder largest ownerA binary variable, 1 indicates that a family is the largest shareholder in the firm and that the founder serves as a >5% owner or officer. Source: Compact Disclosure; firm proxy.
Other family largest ownerA binary variable, 1 indicates that a family is the largest shareholder in the firm and that the founder is no longer present in the firm as a >5% owner or officer. Source: Compact Disclosure; firm proxy.
Lone founder CEOA binary variable, 1 indicates that a founder holds the title of chief executive officer (CEO). Source: Compact Disclosure; firm proxy.
Family CEOA binary variable, 1 indicates that a family member holds the title of chief executive officer (CEO). Source: Compact Disclosure; firm proxy.
Family founder CEOA binary variable, 1 indicates that a family founder holds the title of chief executive officer (CEO). Source: Compact Disclosure; firm proxy.
Other family CEOA binary variable indicating that a family member other than the founder holds the title of chief executive officer (CEO). Source: Compact Disclosure; firm proxy.
Growth/conservation strategySum of standardized scores for R&D, advertising, investment, leverage, and cash holdings minus dividend payout ratio. Conservation is taken as the opposite polarity to growth. Source: Compustat.
Research & developmentResearch and development expenses divided by total sales. Source: Compustat.
AdvertisingAdvertising expenses divided by total sales. Source: Compustat.
InvestmentCapital expenditures divided by plant property and equipment. Source: Compustat.
LeverageThe sum of long term debt plus debt in current liabilities divided by the sum of long term debt plus debt in current liabilities plus the book value of common equity. Source: Compustat.
Cash holdingsCash plus short term investments divided by plant, property, and equipment. Source: Compustat.
Dividends to earnings ratioThe sum of common and preferred dividends divided by operating income before depreciation. Source: Compustat.
Total shareholder returnsTSR represents a firm level market performance measure obtained by compounding each firm's daily market returns in its respective fiscal year. Source: CRSP.
Firm ageThe difference between the year 2000 and the firm's founding year. Source: firm proxy; firm website; Lexus-Nexis; Hoovers.
Log of salesThe natural log of annual net sales. Source: Compustat.
Inside directors ratioThe ratio of inside directors to total directors. Source: firm proxy.
5% ownerThe ownership percentage of all non-family or non-lone-founder block-holders who hold a 5% or greater ownership stake. Created by summing the non-family or non-founder 5% or greater ownership stakes and dividing this by the firm's total shares outstanding. Source: Compact Disclosure; Compustat; firm proxy.
SupersharesA dummy variable set to equal 1 when a firm has a vehicle in place which creates a differential source of power. Example: differential voting over various classes of stock. Source: Compact Disclosure; firm proxy.
Beta (market risk)The average value weighted returns in which the firm's daily returns are regressed against the returns of the overall market. Source: CRSP.
Debt to equity ratioLong term plus short term debt divided by the market value of common equity. Source: Compustat.
Table III. Descriptive statisticsa
Variable Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
  1. Note: a Correlations greater than 0.07 or less than −0.07 are significant at p < 0.05.

 1 Lone founder largest owner0.080.27                
 2 Family largest owner0.170.38−0.13               
 3 Growth/conservation strategy0.182.650.18−0.03              
 4 R&D0.020.04−0.01−0.030.44             
 5 Advertising0.010.020.060.030.470.07            
 6 Investment0.220.170.140.080.600.160.11           
 7 Leverage0.450.30−0.02−0.130.05−0.17−0.06−0.30          
 8 Cash holdings1.233.770.110.000.510.060.020.27−0.19         
 9 Dividends/earnings0.090.11−0.210.01−0.560.00−0.07−0.350.08−0.18        
10 Total shareholder returns0.210.620.060.010.120.06−0.020.13−0.110.17−0.10       
11 Firm age60.1044.27−0.220.06−0.220.110.04−0.280.13−0.140.43−0.09      
12 Log of sales8.021.12−0.10−0.08−0.020.110.13−0.140.10−0.050.23−0.070.28     
13 Inside directors ratio0.240.150.120.200.18−0.070.060.21−0.110.13−0.260.04−0.27−0.25    
14 5% owners0.220.22−0.10−0.200.02−0.03−0.020.000.09−0.11−0.14−0.06−0.09−0.11−0.06   
15 Supershares0.110.310.090.270.06−0.050.100.050.03−0.01−0.040.00−0.05−0.080.10−0.01  
16 Beta0.821.160.080.000.220.120.030.18−0.070.15−0.170.09−0.11−0.040.070.000.00 
17 Debt/equity0.591.030.07−0.080.02−0.18−0.05−0.160.55−0.12−0.06−0.25−0.05−0.020.030.150.10−0.04

For preliminary classification purposes, a firm was assigned to the broad lone founder category if it had a single or several unrelated owners who are founders and also officers, directors, or owners of more than 5 per cent of the shares, with no family members involved. It was assigned to the broad family category where there were multiple members from the same family who are officers, directors or >5% owners, contemporaneously or as descendants (mean family ownership of family firms was almost 20 per cent). We studied only lone founder and family businesses in which the lone founder or the family were (a) the largest shareholders in the company, or (b) served as CEOs to ensure that these parties could have significant impact. For family firms, we also distinguished those firms where the family founder was still present from other family firms. In all instances the comparison sample was other non-family and non-founder firms on the Fortune 1000.

Models

Dependent variables.  Our composite strategy variable is a growth/conservation continuum. Based on the literature on growth and conservation strategies discussed above (i.e. Davidsson et al., 2002; Harrigan and Porter, 1983; Henderson, 1979; Hofer and Schendel, 1978; Kirzner, 1979; Porter, 1980, pp. 267–74), we assessed the composite index according to the following six variables: R&D/sales, advertising/sales, investment (capital expenditures/property, plant, and equipment), financial leverage (debt/debt + equity), cash holdings (cash + liquid assets/property, plant, and equipment), and dividends/earnings. Given the anticipated negative relationship between dividends and the other components, that last variable was subtracted. There were few missing values for most variables, except for advertising and R&D. Here, missing values were coded as 0, as firms were required by law to report these expenditures wherever they were significant. All variables were assessed for skewness and kurtosis, and where necessary were either log transformed and/or their outliers were converted to their respective variable's 99th or 1st percentile, as appropriate. All variables were standardized before being summed for inclusion into the growth/conservation measure.

As strategy composites were derived from prior theory rather than empirically, their statistical properties were evaluated. The inter-item correlation alpha for the growth/conservation composite was 0.51, which is deemed very acceptable for any broad organizational construct with conceptually distinct terms (Van de Ven and Ferry, 1980, pp. 78–81). This composite should be viewed as a summative index rather than a unidimensional construct. However, given the breadth of the composite, we also present summary results for each of its components in Table VI, based on models incorporating the same control variables as employed in Table IV. Performance hypotheses were assessed according to total shareholder returns. The variables are defined in Table II.

Table VI. Robustness checks and summary of results by component variablesa
  (a) Lone founder businesses (b) Family businesses
H# Dependent variables Hypothesized comparisons Big owner CEO H# Hypothesized comparisons Big owner CEO H# Hypothesized comparisons Founder owner Founder CEO
  1. Notes: a Y indicates that the hypothesis is supported, N indicates lack of support. All time-series/cross section models in rows 1 to 8, incorporate the same statistical algorithms, predictors, and control variables as the corresponding models of Tables IV and V. Full results for the TSCS and Heckman models are available from the authors.

   Strategy            
1H1Growth strategyHigherYYH3LowerYYH5AverageYY
2  R&DHigherNN LowerYY AverageYY
3  AdvertisingHigherNN LowerYN AverageYY
4  InvestmentHigherYY LowerNN AverageNY
5  LeverageHigherNN LowerYY AverageYY
6  DividendsLowerYY HigherYY AverageYY
7  CashHigherYY LowerNN AverageYY
   Performance            
8H2TSRHigherYYH4LowerNNH6AverageYY
Table IV. Results of time series cross sectional regression of growth strategy on governance
  Model 1: Model 2: Model 3: Model 4:
H1, H3 H1, H3 H5 H5
Owner CEO Owner CEO
Lone founder vs. family Lone founder vs. family Family founder vs. other family vs. lone founder Family founder vs. other family vs. lone founder
β S.E. β S.E. β S.E. β S.E.
  • Notes: a For parsimony results for 48 separate two digit SICs are suppressed.

  • * 

    p ≤ 0.05;

  • ** 

    p ≤ 0.01;

  • *** 

    p ≤ 0.001.

Lone founder largest owner0.986**(0.350)  1.013**(0.348)  
Family largest owner−0.693***(0.214)      
Family founder largest owner    −0.250(0.339)  
Other family largest owner    −0.924***(0.220)  
Lone founder CEO  0.747**(0.252)  0.777**(0.252)
Family CEO  −0.529**(0.195)    
Family founder CEO      0.301(0.339)
Other family CEO      −0.788***(0.206)
Beta (market risk)0.061*(0.028)0.063**(0.024)0.061*(0.028)0.063**(0.024)
5% owner−0.095(0.383)−0.226(0.338)−0.114(0.381)−0.254(0.338)
Supershares0.360(0.302)0.120(0.273)0.377(0.307)0.191(0.278)
Inside directors ratio1.617**(0.540)1.741***(0.501)1.561**(0.541)1.576***(0.492)
Firm size0.136*(0.062)0.147**(0.059)0.136*(0.062)0.150**(0.058)
Firm age−0.003(0.002)−0.003(0.002)−0.002(0.002)−0.003(0.002)
Industry membershipa
N3061316030613160
Wald chi-sq8783.41***8823.67***8687.81***8309.28***
Number of firms645683645683

Predictor variables.  As predictors we used each of the previously defined dummy indicators of lone founder and family ownership and management. To test Hypotheses 1–4 we employed a dummy to indicate only firms in which the family or the lone founders were the largest shareholders in the firm. To assess robustness of the findings we also tested these hypotheses where the founder or a family member served as CEO. In order to further test robustness we also used family and lone founder share ownership and vote control hurdles of 20 per cent each. To test Hypotheses 5 and 6, we selected only family firms in which the family founder was present.

Control variables.  Strategy may well be a function of the industry the firm is in, as well as its age and size. For example, conservation strategies may be more common in stable industries, among larger firms, and in older businesses (Porter, 1980). For growth strategies, the opposite is apt to be the case (Miller, 1983). Thus each of the models testing strategy Hypotheses 1, 3, and 5 control for these factors: namely industry at the two-digit SIC level, firm age, and the natural log of sales. It is particularly important to control for age and size as these aspects of the corporate life cycle may influence a firm's growth rate. The models controlled as well for governance factors. These included the presence of inside directors and of major (5%) non-family or non-lone founder block-holders, either of whom might bring a business- as opposed to a family-perspective to the board (Shleifer and Vishny, 1997). We also controlled for the use of special voting shares (‘supershares’) which augment family or lone founder control without corresponding ownership (Maury, 2006; Morck et al., 2005). Finally, we incorporated firm beta – or market risk, often associated with growth orientations. Following Anderson and Reeb (2003) and Villalonga and Amit (2006), our models assessing performance Hypotheses 2, 4, and 6 control for the above variables as well as two others: investment (capital expenditures/property, plant, and equipment), and debt/equity.

Analyses and Robustness

Tables IV and V present our time-series, cross-sectional findings, reporting results from generalized estimating equation models (Liang and Zeger, 1986; Zeger et al., 1988). These panel models were especially appropriate given the temporal stability of the predictor dummies (Liang and Zeger, 1986). They also incorporate Huber–White clustered standard errors to control for unobserved firm fixed effects and adjust for firm-specific autocorrelation (Peterson, 2009). As the models are not maximum likelihood, chi-square is the only available goodness of fit measure. All models are statistically significant at beyond the 0.001 level. Also, wherever the hypothesized predictors are significant, the full models also show statistically significant increases in chi-square over nested models without the hypothesized predictors.

Table V. Results of time series cross sectional regression of performance (TSR) on governance
  Model 1: Model 2: Model 3: Model 4:
H2, H4 H2, H4 H6 H6
Owner CEO Owner CEO
Lone founder vs. family Lone founder vs. family Family founder vs. other family vs. lone founder Family founder vs. other family vs. lone founder
β S.E. β S.E. β S.E. β S.E.
  • Notes: a For parsimony results for 48 separate two digit SICs are suppressed.

  • + 

    p ≤ 0.10;

  • * 

    p ≤ 0.05;

  • ** 

    p ≤ 0.01;

  • *** 

    p ≤ 0.001.

Lone founder largest owner0.159**(0.052)  0.162(0.052)  
Family largest owner0.009(0.027)      
Family founder largest owner    0.045(0.051)  
Other family largest owner    −0.010(0.0.31)  
Lone founder CEO  0.171***(0.043)  0.175***(0.043)
Family CEO  0.005(0.027)    
Family founder CEO      0.106+(0.056)
Other family CEO      −0.032(0.028)
Beta (market risk)0.019(0.017)0.024(0.015)0.018(0.017)0.023(0.015)
5% owner0.062(0.042)0.034(0.041)0.060(0.042)0.034(0.040)
Supershares0.025(0.036)−0.002(0.038)0.026(0.037)0.010(0.039)
Inside directors ratio−0.004(0.081)−0.091(0.074)−0.010(0.081)−0.104(0.073)
Firm size−0.015(0.010)−0.025**(0.010)−0.015(0.010)−0.024**(0.010)
Firm age−0.001**(0.000)−0.001*(0.000)0.000*(0.000)−0.001*(0.000)
Investment0.195(0.173)0.140(0.153)0.186(0.176)0.136(0.153)
Debt to equity−0.150***(0.010)−0.154***(0.011)−0.150***(0.010)−0.155***(0.011)
Industry membershipa        
N3153324331533243
Wald chi-sq2877.97***3043.40***2941.80***2387.83***
Number of firms662698662698

To establish the robustness of the time series findings of Tables IV and V, we ran parallel OLS models using five-year averages of the same variables. Fixed effects models were inappropriate given the stability of our governance predictors over the five-year period of analysis. The OLS models (available from the authors) incorporate only firms whose ownership status did not change over the five years of data. Where hypothesized coefficients were significant in the predicted direction for the panel analyses of Tables IV and V, all were also significant in the predicted directions for the OLS analyses. Results were also confirmed when we used family and lone founder ownership and vote control hurdles of over 20 per cent, and when we stipulated major ownership and serving as CEO as joint role criteria.

Survival bias and endogeneity may arise in examining the relationships between governance and performance. Not only may governance influence performance, good or poor performance might also cause a change in governance – the sale by a family or lone founder of the business, for example. Thus, following Greene (2003, pp. 787–90), we checked and controlled for endogeneity in our analyses using Heckman two-step treatment effect regressions for all our indicators of lone founder- and family-ownership and management. The first stage of the procedure is a probit analysis that regresses the firm governance dummies against variables that distinguish among lone founder, family, and other businesses. Following Miller et al. (2007) and Villalonga and Amit (2006), these predictor variables include supershares, firm age, sales growth, debt to equity, two-digit SIC dummies, and the average age of directors. To establish robustness we varied the predictors for the probit analyses by dropping director age, growth, and beta. This did not significantly change the results. The second stage of the Heckman procedure regresses our dependent variables on the predicted values from the first stage, adjusting all errors for the two-stage estimation process. The Heckman results confirmed those of Tables IV and V, except that in the latter, family founder CEOs showed performance that was significant at beyond the 0.05 instead of beyond the 0.10 level (the detailed analyses are available from the authors).

In order to further check for sample selection bias we gathered data from a random sample of 100 smaller, non-Fortune 1000 companies. The random sample was taken from the Compustat database for the year 2000 from a universe of almost 5000 publicly traded firms. Random numbers were used to select 100 of these firms. Means were calculated for the strategy and performance variables for lone founder, family, and other businesses, and means comparisons were run against the corresponding categories of the Fortune 1000 (the exact findings are available from the authors). An important concern regarding a Fortune 1000 sample is that these firms have been unusually successful, and that, at least the lone founder firms, which tend to be younger, have grown especially rapidly. We found that the differences between the Random 100 sample and the Fortune 1000 in performance and strategy means were mostly non-significant for lone founder, family, and other categories. Specifically, the lone founder, family, and other categories did not differ significantly between these samples along total shareholder returns. For the growth/conservation strategy dimension, there is no significant difference for family or lone founder businesses between the samples. These findings help to allay the general concern that the Fortune 1000 sample is unrepresentative, and the fear that the lone founder category of the Fortune 1000 shows an inflated tendency towards growth. This is not to say, however, that readers should take our results as applying to smaller companies. Further research would be needed to establish that.

A number of further measures were employed to establish the robustness of our findings. Multiple indicators for ownership and management were investigated. For example, we ran all analyses for the Chairman and CEO-Chairman positions using the same control variables as on Table IV. But these findings were so close to the CEO results we report that we refrain from presenting them. We also systematically varied our sets of control variables: specifically, beta, and supershares individually were added to/deleted from our models with no material changes in the findings for the hypotheses.

Finally, we ran models for each individual component of our strategy composite, using the same control variables as in Table IV. The results are summarized in Table VI, and show significant convergence. In summary, the majority our findings were consistent across different sets control and component variables, and across different types of analyses (cross-sectional, panel, and treatment regressions). Any material differences are reported in our results and discussed.

FINDINGS

Tables IV and V present our findings for strategy and performance, respectively. The coefficients reported are from time-series cross-section regressions (random effects models with robust standard errors). The two-digit industry SIC dummy variables were suppressed to save space.

Hypotheses 1 and 2.  It was hypothesized that, following an entrepreneurial logic, businesses in which lone founders were the largest shareholders would adopt growth strategies and outperform. Tables IV and V confirm that finding – specifically they show founder owned firms in which the founder is the largest shareholder to be more apt to embrace growth strategies than other Fortune 1000 firms (Model 1 of Table IV) and also to outperform in shareholder returns (Model 1 of Table V). When we relaxed the largest shareholder requirement, and stipulated that a lone founder serve as CEO, the findings again were confirmed (Model 2 of Tables IV and V). In examining the disaggregated components of the growth strategy (Table VIa), it appears that lone founder firms do indeed invest more generously in their infrastructures; they also pay less dividends and accumulate more cash. They do not, however, engage in more R&D, nor employ more leverage than the average Fortune 1000 firm. Their growth orientation, therefore, appears to avoid these more risky initiatives.

Hypotheses 3 and 4.  Hypothesis 3 stipulated that, following a family logic, firms in which the family was the largest shareholder would be more prone than other firms to pursue conservation strategies (i.e. less apt to pursue growth strategies). Model 1 of Table IV confirms that, and Model 2 of the table confirms that the hypothesis is also supported when a family member serves as CEO. Table VIb suggests that unusually generous dividends, reduced financial leverage, and below-average R&D and advertising expenditures were unusually common among family firms. Low levels of cash and investment were less in evidence. Overall, then, family businesses favour conservation over growth.

Hypothesis 4 stipulated that firms in which a family was the largest owner would underperform in returns to shareholders. This hypothesis was not supported for the family as major owner condition, nor when a family member serves as CEO. Specifically, Models 1 and 2 of Table V show that family businesses do not underperform in shareholder returns, nor do they outperform. The reasons for sustaining at least average performance may, in part, be ascribed to the advantages of closely held companies where influential family owners are better able to control managerial opportunism – or engage in more functional forms of conservatism.

Hypotheses 5 and 6.  Firms run by family founders were expected to be subject to both entrepreneurial and familial logics, and hence Hypotheses 5 and 6 stipulated that these firms would fall between lone founder owned and family owned enterprises in their strategic and performance orientations. Models 3 and 4 of Table IV indicate that the firms of family founder-owners or family founder-CEOs do not differ from other Fortune 1000 firms in their growth orientations. That is, they fall between lone founder and other family firms in that regard. However, according to Table V, Model 4, when the family founder stays on as CEO, performance is superior – albeit only at the 0.10 level (0.05 in the Heckman analyses). The founder effect appears to be of some significance, not only in lone founder but family founder businesses. When the founder is no longer CEO, the performance advantage disappears, and strategies once again are those of conservation (see ‘other family’ rows in Model 4, Tables IV and V).

The results suggest that the effects of major ownership and of CEO roles were largely comparable. The logics embraced, in other words, might well accrue under both major ownership and management positions. Founder presence also seems to have some impact, as the departure of the founder CEO in family firms appears to presage a transition towards familial as opposed to entrepreneurial logics – and hence towards conservation strategies (Table IV, Model 4). Perhaps this is because post-founder family CEOs have never acted as entrepreneurs or business builders, and often serve at the will of the family, and hence the entrepreneurial logic does not apply (Gersick et al., 1997; Ward, 2004).

DISCUSSION AND CONCLUSION

Governance

The literature on corporate governance has for many years debated the impact on performance of different levels of ownership concentration and of owner-management (Shleifer and Vishny, 1997). Using an agency framework, it has argued that concentrated ownership and owner-management contribute positively to performance. The theme of that work has been the ability of powerful owners to reduce agency costs in their pursuit of economic self-interest (Jensen and Meckling, 1976). Unfortunately, these analyses have produced conflicting empirical findings. Some studies suggest superior returns from concentration, others do not (see the reviews by Morck et al., 2005 and Shleifer and Vishny, 1997). Indeed, even within the class of public family enterprises there is significant debate about whether these concentrated firms outperform – or underperform (compare Anderson and Reeb, 2003; Maury, 2006; Miller et al., 2007; Pérez-González, 2006; Villalonga and Amit, 2006). More recent work on family firms has indicated that different types of owners and executives may perform differently – for example, lone founder firms experience higher market valuations than family and family founder firms (Miller et al., 2007). But performance is measured purely according to market valuation, strategy is ignored, and nothing is said to resolve discrepancies across studies.

Our thesis was that these discrepancies were due to the failure to distinguish among the social contexts of different types of owners and executives, and the effects these contexts might have on the role identities and institutional logics of those parties. We argued that such distinctions might explain differences in strategy and performance. Specifically, family owners and CEOs were expected to adopt family nurturing identities and familial logics; lone founders would adopt entrepreneurial identities and logics. The consequences would be conservation strategies and mediocre returns for family firms, and growth strategies and out-performance for lone founder firms. Family founders were argued to mix these identities, logics, and outcomes. These expectations were largely borne out by our research.

We found that lone founder businesses pursued a strategy of growth and earned superior shareholder returns. These findings obtained when the founder was the biggest owner or the CEO. By contrast, family businesses pursued a conservation strategy and garnered average total returns for shareholders. Again, these findings emerged whether the family member was the largest owner or the CEO. A strategy that fell between the polarities of growth and conservation pertained to family-founders, who marginally did outperform, albeit not to the extent of lone founders. Clearly, it is time to go beyond economic models of self-interest and agency and to employ richer social conceptions of actors playing key governance roles.

Our findings that family businesses do not outperform contradict earlier studies: those that claim family businesses do outperform (Anderson and Reeb, 2003; McConaughy et al., 1998; Villalonga and Amit, 2006), and those that claim they underperform (Bennedsen et al., 2007; Bloom and Van Reenen, 2007). Again, those studies had adopted an agency logic that fails to distinguish family founder firms from lone founder firms. We have argued for the importance of such distinctions as they demarcate a clear contrast in social context. As noted, and consistent with their respective entrepreneurial and familial orientations, only lone founder firms outperform; family firms do not.

Our finding that family firms do not significantly underperform also challenges scholars condemning family business nepotism and entrenchment (e.g. Claessens et al., 2002; Pérez-González, 2006). Families do tend to pursue conservation strategies, but do not notably penalize their public shareholders as a result, at least not vis-à-vis non-lone founder firms. A philosophy of nurturing the family apparently does not jeopardize the viability of the firm.

Our findings suggest that family involvement in a public family business may at times be limiting. If family demands are the challenges this research suggests they are, it will be useful for enterprises to limit family involvement to relatives who place a business agenda above that of a family agenda, and have sufficient expertise to master their jobs. Unfortunately, the painful issue of pruning unproductive or problematic family members from a firm has been neglected by scholars. Yet pruning and instilling meritocracy may be a key priority, especially in competitive environments in which conservation strategies will almost guarantee stagnation. Useful remedies are indicated by some European family firms that have their family shareholders elect one or two expert representatives to serve on the board. Then all family shareholders vote every few years for the representatives, but otherwise have little involvement with the company (Landes, 2006). Firms also can hire top executives from outside the family, especially after the founder has gone (Bennedsen et al., 2007).

Our choice of the Fortune 1000 sample was notable. It is a useful sphere within which to study the blending of family and market logics because it has been deemed to be so dominated by a quintessentially market regime (Morck et al., 2005). Whereas it would come as no surprise to detect manifestations of family logic in small firms owned, managed, and populated by families, that is less true of major corporations with significant public ownership and ‘professional’ management. Indeed, investor safeguards to enforce market discipline within American stock exchanges are said to be among the best in the world (Shleifer and Wolfenzon, 2002). There, then, is where the market logic should reign – especially given today's obsession with quarterly earnings and shareholder protections (Jacobs, 1996). Thus any manifestations of a familial logic that qualifies or even overrides market logic in such a context is especially significant.

Institutional Logics

Our research sheds light on a central debate in the corporate governance literature. But it also extends the application of institutional logics to that domain. Moreover, it stands out from most of the previous work on logics in several ways. First, the study of logics has centred mostly on the institutions of capitalism and the professions, and underplayed the role of the family. Yet in family firms the family institution – and the familial logic to which it may give rise in particular contexts, can be decisive in determining actors' role identities, sources of legitimacy and authority, and values and missions – even in large, mature public companies (Chung and Luo, 2008; Greenwood et al., 2010).

Earlier work on logics has been concentrated at the level of organizational fields such as professional service firms (Greenwood and Suddaby, 2006), healthcare organizations (Galvin, 2002; Reay and Hinings, 2005), and publishing houses (Thornton, 2004), and it has focused on logics adopted at a more impersonal organizational level. By contrast, our study encompasses a wider swath of firms in many different industries, and distinguishes them on the basis of broad governance regimes that span many fields and impinge quite directly on the social context of major personal owners or executives. Thus the emphasis is on societal logics that take effect at the level of the individual actor. This takes us back to the original emphasis of Friedland and Alford (1991).

It would be unwise to conclude from the results of this study that the market institution does not impact family owners, or that the institution of the family does not shape the behaviour of even the most entrepreneurial founders. Individuals' identities and their logics are influenced by multiple institutions, and moulded in various degrees by each of them. The institutional factors impinging on an individual via roles, values, and structural and historical circumstance will determine the balance of logics reflected in behaviour.

Limitations and Future Research Directions

Methodological issues.  There are a number of limitations to the current research. First, any study of public corporations is biased. It does not reflect the behaviour of private companies, and many family businesses are just that. It is also true that firms that have made it onto the list of the Fortune 1000 are special: already they have in some sense outperformed companies that have not made it onto the list. Thus care must be taken not to generalize our findings beyond large US publicly traded firms. In particular, our conclusions may not apply to smaller, younger, and private companies. Where families wholly own and run smaller businesses, their association with the business may be quite intimate, thereby fostering a nurturing attitude that husbands resources more carefully and tilts the strategic balance towards growth (Miller and Le Breton-Miller, 2005). There is also the issue of survival bias, for which our methods can only imperfectly control.

Our measures of ownership and family involvement were limited. First, there may exist partners and family owners who possess fewer than 5 per cent of the shares of a firm and do not serve as officers or directors. Given statutory reporting requirements in the proxies, those individuals were invisible to us and so our estimates of family ownership are almost certainly low. Also, our measures of strategy were truncated as we did not have available to us information on goals, culture, human resources policies, and functional orientations that would provide a more complete picture of strategic conduct. A more focussed field study might address these gaps.

Conceptual issues.  We cannot prove that our findings are a direct result of contrasting social contexts, identities, or logics. Unfortunately, like many studies, our research suffers from an inability to discern the micro-processes that cause contexts to mould role identities and logics (Thornton and Ocasio, 2008, p. 120). More work needs to be done to identify the cognitive, normative, and political forces that influence how roles and logics are adopted. Identity theory and social identity theory may offer important insights here – insights that require a more fine-grained study of social interaction, and the processes driving role identification and role salience (Burke, 1980; Burke and Reitzes, 1981; Hogg et al., 1995; Stryker, 1980). In the family context, long term emotional connections with family members inside and outside the business may have an important impact on the worldviews, loyalties and priorities of family members. But it would be useful to study just how and why this happens – to what degree does it rely on commonality of background and education, frequency of interaction, family harmony and closeness, personality, or purposeful socialization? The question of how and why lone founders assume their identities and logics is just as complex. To what extent is the entrepreneurial role and logic embraced as a function of career background, projective identification, or associations with investors and venture capitalists? And are these entrepreneurial role identifications or ‘self-categorizations’ with a more abstract, physically remote group (the set of entrepreneurs) more driven by social attraction than social interaction (Hogg et al., 1995, p. 263). These are all issues that require study to establish the fine grained details of our broad arguments.

ACKNOWLEDGMENTS

The authors are grateful to Professor Ken Craddock for his useful contributions. They would also like to thank Sadia Chowdhury, Luc Farinas, Tim Holcomb, Carla Jones, and Soranna Pramualmitra for their help with the data gathering. They are pleased to acknowledge support from the Social Sciences and Humanities Research Council of Canada.

NOTE

  • [1] 

    Agency theory concerns the power to optimize self-interest, and it assumes economic self-interest as the primary motivator. As such it ignores that actors may define interest in different ways (Gomez-Mejia and Wiseman, 2007; Schulze et al., 2001). Notably, it ignores the fact that ‘interest’ is in part socially defined, and thus that actors may be motivated differently because of their structural and social positions.

Ancillary