Abstract
Calling upon stakeholder theory and the socioemotional wealth (SEW) literature, we investigate how SEW impacts the decline-stemming strategies of family firms. Drawing on a recent conceptualization of SEW, we validate a two-dimensional measurement of the construct using a content analysis technique. Our empirical test on a sample of publicly traded family firms in need of turnaround suggests that the strategic preferences of family firms change depending upon the type of SEW (extended vs. restricted) the owning family values the most. The fine-grained characterization of SEW adopted in this study accounts for within-family-firm differences and thus enables the reconciliation of conflicting findings in the literature.
Keywords
Scholars have devoted considerable effort to understand turnaround behavior and how firms respond when facing a prolonged and significant decline in financial performance (see Trahms, Ndofor, & Sirmon, 2013, for a comprehensive review). While the erosion of stakeholder support has emerged as a salient issue (Arogyaswamy, Barker, & Yasai-Ardekani, 1995; Trahms et al., 2013), stakeholders’ concerns can be ameliorated by developing strategies that halt a firm’s downward spiral in performance and gathering the resources necessary for its recovery (Arogyaswamy et al., 1995). These decline-stemming strategies are critical because they enable the firm to build a foundation for its recovery by addressing the dysfunctional consequences of performance decline (Pearce & Robbins, 1993). However, the role stakeholders play in the turnaround process in family businesses has been partially overlooked (Decker, 2016; Trahms et al., 2013).
Evidence about the first response to a turnaround situation in family-owned businesses (FOBs) provides mixed arguments with conflicting results. Some arguments in the literature suggest that FOBs with high family ownership respond to drops in financial performance that threaten firm survival with significant corporate restructuring (Kavadis & Castañer, 2015). Others (e.g., Block, 2010; Essen, Strike, Carney, & Sapp, 2015) contend that the desire of FOBs to be socially responsible toward employees and to defend their reputation hinders significant responses to a turnaround situation. This inconsistency may be due to the fact that FOBs are a highly heterogeneous group (Salvato & Aldrich, 2012), pursuing different combinations of both financial and nonfinancial goals (e.g., Gómez-Mejía, Haynes, Núñez-Nickel, Jacobson, & Moyano-Fuentes, 2007; Gómez-Mejía, Patel, & Zellweger, 2018). As a result, the fulfillment of a stakeholder’s needs may vary depending upon which nonfinancial resource the owning family aims to receive (Vardaman & Gondo, 2014). Previous studies also grouped all nonfinancial objectives an owning family expects to derive from the firm (Berrone, Cruz, & Gómez-Mejía, 2012) under the all-encompassing and unidimensional construct of socioemotional wealth (SEW) (e.g., Kavadis & Castañer, 2015). As SEW comprises multiple nonfinancial benefits (Berrone et al., 2012; Debicki, Kellermanns, Chrisman, Pearson, & Spencer, 2016), the intricate relations between financial and nonfinancial goals and different forms of SEW can only be disentangled if we account for its multidimensionality (Vardaman & Gondo, 2014).
We bridge the above knowledge gaps by addressing the following research question: How does SEW affect the choice of decline-stemming strategies in FOBs facing a turnaround situation? We do this by incorporating insights from stakeholder theory and the SEW mixed-gamble literature. From a stakeholder theory perspective, we argue that the owning family facing a turnaround situation accommodates the claims of stakeholder groups who provide critical resources (Decker, 2016). From an SEW perspective, we recognize that FOBs pursue both financial and nonfinancial (i.e., socioemotional) benefits (Cruz & Justo, 2017; Gómez-Mejía, Neacsu, & Martin, 2017; Gómez-Mejía et al., 2018). The nonfinancial resources an owning family desires may change depending upon which type of SEW (i.e., restricted or extended SEW) it values the most (Cruz, Larraza-Kintana, Garcés-Galdeano, & Berrone, 2014; Vardaman & Gondo, 2014). On the one hand, an owning family who advances the achievement of a family-centered agenda (i.e., restricted SEW, or SEWr) will prioritize the interest(s) of family over those of nonfamily stakeholders (Miller & Le Breton-Miller, 2014). These FOBs avoid significant decline-stemming strategies in order to protect the interests of providers of financial resources that are instrumental in preserving the family’s relatively unconstrained control over the firms. In contrast, FOBs that aim to preserve a positive family image and reputation among stakeholders, long-lasting relationships with nonfamily stakeholders, and the long-term well-being of the family (i.e., extended SEW, or SEWe) deploy significant decline-stemming strategies to protect this form of SEW.
We test our hypotheses using original data from a sample of publicly traded family firms that faced turnaround situations between 2000 and 2012. Overall, this study contributes to the literature in multiple ways. First, we make a theoretical contribution in that we address the shortcomings of stakeholder theory when used in isolation to explain the responses of FOBs in turnaround situations. Due to the idiosyncratic nature of FOBs, this established organizational theory cannot be directly applied to the family business context. Hence, we develop a theoretical argument that brings together stakeholder theory with knowledge advanced in the SEW literature (Berrone et al., 2012; Miller & Le Breton-Miller, 2014). Second, as a result of its enhanced explanatory power, our study reconciles conflicting evidence concerning the first response of family firms to a turnaround situation (Faghfouri, Kraiczy, Hack, & Kellermanns, 2015; Kavadis & Castañer, 2015). Our multidimensional, yet parsimonious conceptualization of SEW provides a finer-grained explanation for the strategic preference of family firms (Cater & Schwab, 2008).
By adopting a multidimensional SEW construct, this study can better explain heterogeneity across family firms (Chua, Chrisman, & De Massis, 2015; Salvato & Aldrich, 2012). Our approach also lends itself well to the investigation of strategic preferences of FOBs beyond decline-stemming strategies. As the first study to validate a two-dimensional conceptualization and operationalization of SEW (SEWr and SEWe) through computer-aided text analysis (CATA), this reconceptualization of SEW makes a theoretical contribution (Colquitt & Zapata-Phelan, 2007) as it assists scholars to theorize with higher precision about the strategic preferences of family businesses.
Strategic Responses to a Turnaround Situation
A number of studies address the implications of performance below aspirational levels and performance gaps on FOBs’ strategic preference (e.g., Gómez-Mejía et al., 2018; Patel & Chrisman, 2014). Yet comparatively fewer studies directly investigate turnaround situations, or extended decline in financial performance when the firm’s survival is at risk. Indeed, only a handful of studies attempt to disentangle the relationship between family firm variables and turnaround strategies (Cater & Schwab, 2008; Chirico, Salvato, Byrne, Akhter, & Arriaga Múzquiz, 2018). Our study aims to bridge this gap.
Recent systematic reviews of turnaround literature indicate that firms are in need of turnaround when they experience at least 3 years of consecutive decline in performance following a period of growth (Robbins & Pearce, 1992). To halt the performance decline and return to growth, firms engage in a two-stage turnaround process (Pearce & Robbins, 1993; Trahms et al., 2013). The first stage (decline stemming) includes actions aiming to generate higher efficiencies and re-focus the business. This stage typically includes: (a) aggressive cost-cutting (cost retrenchment); (b) severe asset cuts (asset retrenchment); and (c) intense focus on sales improvement either through price reduction or through investment in marketing funded via cuts in all other functions of the firm (Arogyaswamy et al., 1995). The second stage (recovery) involves investment in repositioning the firm and recovery of financial performance (Pearce & Robbins, 1993; Trahms et al., 2013).
Our focus is on the critical first step in responding to a turnaround situation as decline-stemming strategies are a crucial element for a successful turnaround (Arogyaswamy et al., 1995; Schmitt & Raisch, 2013). During turnaround situations, stakeholders become overly protective of their stake in the declining firm and may in turn influence managerial decisions to preserve their self-interest (Arogyaswamy et al., 1995; D’Aveni, 1989). Decline-stemming strategies address the consequences of decline, including the erosion of stakeholder support (Barker & Mone, 1994). Therefore, firms need to rely on decline-stemming strategies to preserve the stakeholder support needed for the development of recovery strategies.
We argue that each of the strategies preserves the interest of a specific category of stakeholders. Thus, we expect the firm’s preference for a specific strategy to depend on both the pressure exerted by a specific stakeholder category and the relevance ascribed by the firm to those stakeholders (Jawahar & Mclaughlin, 2001). Therefore, stakeholder theory can help explain and predict the turnaround strategic preferences of family firms.
Theoretical Framework and Hypotheses
This study captures the heterogeneity of FOBs by investigating the role each SEW benefit plays in their strategic behavior during turnaround. To address this, we bring together recent theoretical insights from the SEW literature along with stakeholder theory prescriptions concerning the likely decline-stemming strategies pursued by a family firm.
Stakeholder theory suggests that managers must account for the interests of both internal (e.g., employees, shareholders, governance) as well as external stakeholders (e.g., local communities, the environment, suppliers of critical resources) to successfully run their organizations (Freeman, 1984; Trahms et al., 2013). However, firms in turnaround situations have limited resources and are unable to attend to the claims of all stakeholders (Jawahar & Mclaughlin, 2001). Hence, stakeholders should be managed strategically (Frooman, 1999), especially since decline-stemming strategies are highly likely to accommodate the claims of those stakeholders who control the firm’s access to financial resources (Jawahar & Mclaughlin, 2001). In this context, salient stakeholders are those who control the resources needed to halt the decline, and whose claims are time-sensitive (Donaldson & Preston, 1995). Delaying the fulfillment of these claim(s) might accelerate performance decline and withdrawal of stakeholder support (Tangpong, Abebe, & Li, 2015). Meanwhile, the firm may take a more assertive position toward the claims of less salient stakeholders (Jawahar & Mclaughlin, 2001), who do not control resources crucial to the firm’s survival. While insightful, such an approach suffers from two main shortcomings, especially in the context of FOBs facing a turnaround situation. First, within FOBs, the family represents a powerful and complex stakeholder group with heterogeneous sets of noneconomic goals (Zellweger & Nason, 2008). FOBs, therefore, manage stakeholders as they pursue both financial and SEW benefits (Cruz & Justo, 2017). Second, the definition of important stakeholders may differ across FOBs depending upon the value ascribed by the family decision maker to various nonfinancial benefits (Cennamo, Berrone, Cruz, & Gómez-Mejía, 2012). To address these shortcomings, stakeholder theory needs to be informed by recent advances from SEW research.
Socioemotional Wealth Perspective and Construct Dimensions
SEW consists of “non-financial aspects of the firm that meet the family’s affective needs, such as maintaining its identity, exercising influence, and perpetuating the family dynasty” (Gómez-Mejía et al., 2007, p. 106). Recent advances in SEW research (Cruz & Justo, 2017; Gómez-Mejía et al., 2018) adapted the concept of “mixed gamble” (Martin, Gómez-Mejía, & Wiseman, 2013), which originated in behavioral studies (Bromiley, 2009), to explain strategic decisions of FOBs. According to this perspective, decision-making in family firms results from an assessment of expected gains and losses of both financial and socioemotional wealth (Cruz & Justo, 2017; Gómez-Mejía et al., 2018). However, although insightful, this approach does not account for the multidimensionality of SEW (Berrone et al., 2012; Hauck, Suess-Reyes, Beck, Prügl, & Frank, 2016) or the varied goals and aspirations found across owning families (Chrisman & Patel, 2012). As a result of this heterogeneity, FOBs may balance tensions between financial wealth and SEW differently, and this may also vary across distinct types of SEW (Vardaman & Gondo, 2014).
The multidimensional nature of SEW can be captured by the five-dimension FIBER conceptualization (Berrone et al., 2012). This encompasses family control and influence over the firm (F), family members’ identification with the firm (I), binding social ties (B), emotional attachment (E), and renewal of family bonds to the firm through dynastic succession (R) (see the Appendix [Supplemental Material online] for a detailed description of each dimension). However, significant theoretical and empirical overlap among some of the SEW dimensions limit the utility of the original FIBER model (e.g., Debicki et al., 2016). Additionally, challenges connected with the measurement of the FIBER dimensions hinder scholars’ ability to show causal links between individual FIBER dimensions and relevant outcomes (Miller & Le Breton-Miller, 2014). Although recent efforts validate measures for the FIBER dimensions (e.g., Hauck et al., 2016; Hsueh, 2016), they either exclude the F dimension or employ survey items that fail to address the affective value an owning family may derive from exercising influence and control (Hauck et al., 2016). Therefore, scholars have called for a more parsimonious categorization of SEW typology.
To heed this call, we draw from Miller and Le Breton-Miller’s (2014) dichotomization of SEW in SEWr and SEWe. Restricted SEW includes the emotional benefits resulting from the ability to exercise influence over the firm to achieve highly family-centered objectives (Miller & Le Breton-Miller, 2014). This is similar to the F dimension in the Berrone et al.’s (2012) model. An owning family pursuing SEWr will consider the immediate family as the most important stakeholder group (Miller & Le Breton-Miller, 2014); thus, the short-term well-being (e.g., stock of SEW) of family members drives the strategic posture of these firms (Chrisman & Patel, 2012; Chua et al., 2015). SEWr also shares priorities typical of owning families pursuing the E dimensions of SEW. Indeed, scholars further argue that owning families who pursue SEWr have as priorities the preservation of family control and influence over the firm through a family-dominated management or board (Berrone et al., 2012), the ability to impose unrestrained will over the firm or influence its management, the ability to rely on kinship ties for promotion and hiring decisions (Chua, Chrisman, & Bergiel, 2009; Gómez-Mejía et al., 2007), and the ability to behave altruistically toward family members. Thus, the F and E dimensions of SEW can be aggregated in the higher order SEWr dimension.
On the other hand, SEWe encompasses the emotional benefits of fulfilling the needs of the family across generations as well as those of external stakeholders (Miller & Le Breton-Miller, 2014). It captures items such as: (a) the enjoyment of maintaining good relationships with stakeholders and the creation of long-lasting relationships with partners in order to increase the chances of firm survival, which are critical components of the B dimension of SEW (Berrone et al., 2012); (b) the emotional value of ensuring firm survival across generations, an important element of the R dimension of SEW (Berrone et al., 2012); and (c) the affective value of enhancing a family’s reputation with stakeholders and ensuring an abundance of goodwill toward the family and its business, which are elements of the I dimension of SEW (Cennamo et al., 2012). An owning family pursuing SEWe will consider the family over time and nonfamily stakeholders as salient groups (Miller & Le Breton-Miller, 2014). Thus, the long-term sustainability (e.g., flow of SEW) of the business drives the firm’s strategic posture (Chua et al., 2015). Hence, the I, B and R dimensions form the higher order construct of SEWe.
Extended SEW, Restricted SEW, and Decline-Stemming Strategies
Considered together, stakeholder theory along with SEWr and SEWe help us account for the differences in the way FOBs respond in turnaround situations. Specifically, we anticipate that family firms will concentrate available resources on developing decline-stemming strategies that preserve the support of stakeholders providing access to critical financial and socioemotional (SEWe and SEWr) resources (Cruz et al., 2014; Frooman, 1999). However, the exact type of SEW firms strive to access differs across FOBs. Accordingly, stakeholder groups perceived as important by an owning family that values SEWe may be different from those considered important for SEWr (Miller & Le Breton-Miller, 2014). Therefore, the mixed-gamble logic employed by scholars to explain family firm strategic behavior is further constrained by the value ascribed to different types of SEW.
Table 1 details the likely implications of different decline-stemming strategies for the focal stakeholders in FOBs wherein the owning family values SEWr or SEWe. As shown in Table 1, the literature suggests that owning families who pursue SEWr prioritize the achievement of highly family-centric goals (Chua et al., 2015) such as control and influence over the firm and the family’s current well-being (Miller & Le Breton-Miller, 2014). Thus, from a stakeholder theory perspective, firms will strive to accommodate the claims of the immediate family and of focal stakeholders providing financial resources (SEWr stakeholders) (Jawahar & Mclaughlin, 2001; Miller & Le Breton-Miller, 2014). Specifically, in these firms, the owning family is a dominant coalition in the business; hence, it represents a focal stakeholder. Providers of financial resources (e.g., lenders) are also considered focal stakeholders as they can significantly restrict an owning family’s ability to make discretionary use of the firm’s resources and can interfere with the family’s control and influence over the firm (Abebe & Tangpong, 2018; Decker, 2016), which are important sources of SEWr.
Socioemotional Wealth Implications of Decline-Stemming Strategies.
Note. (+) denotes outcome of decline-stemming strategies consistent with a stakeholder’s expected claim; (−) denotes outcome of decline-stemming strategies contrary to a stakeholder’s expected claim.
Focal SEWr stakeholders are expected to oppose decline-stemming strategies (see Table 1) as they erode the short-term interests of these stakeholder categories (Decker, 2016). Specifically, providers of financial resources may respond negatively to decline-stemming strategies (see Table 1, row 2) because these are seldom cash-generating in the short term and, in some cases (e.g., the focus on improvement in sales), they can be cash-consuming (Decker, 2016; Lai & Sudarsanam, 1997). Hence, to protect their interests, providers of financial resources may withdraw their support to the firm (Tangpong et al., 2015) and indirectly diminish the stock of SEWr in the short term (i.e., by focusing on current SEW) (Chrisman & Patel, 2012). In addition, decline-stemming strategies can inflict direct SEWr loss to the immediate family (Table 1, row 1).
From an SEW mixed-gamble perspective, we posit that decline-stemming strategies are associated with highly probable and immediate losses of SEWr stock (i.e., such strategies increase the fear of diminished SEWr in the present) (Chrisman & Patel, 2012) and highly uncertain potential of SEW gains. Due to the focus of SEWr-oriented FOBs on the current well-being of the family, we anticipate that the owning family will weigh the former more than the latter. Additionally, avoiding intense decline-stemming strategies accommodates the claims of providers of financial resources. Preserving their support helps the owning family weather the firm’s decline temporarily while preserving SEWr in the immediate term (Arogyaswamy et al., 1995). Similarly, in terms of financial wealth, decline-stemming strategies are connected with fairly certain short-term losses (e.g., when the sale of assets is used to reduce debt, value is transferred from the family owner to a creditor) and highly uncertain future gains (there is no guarantee that decline stemming will indeed lead to turnaround) (Filatotchev & Toms, 2006). Therefore, we expect SEWr-oriented FOBs to be reluctant to engage in significant decline-stemming strategies. More formally put:
As shown in Table 1 (rows 4–8), the literature suggests that an owning family that values extended SEW prioritizes the preservation of long-lasting and stable relationships with stakeholders (internal and external) as a way of guaranteeing the well-being of the family over time and the achievement of transgenerational succession (Miller & Le Breton-Miller, 2014). In other words, an owning family that values SEWe is concerned with possible gains or losses of SEW over time (i.e., flow of SEW) (Chrisman & Patel, 2012; Chua et al., 2015). As a result, from a stakeholder theory perspective, an owning family will consider external and internal stakeholders (SEWe stakeholders), including the family itself, as focal.
The tendency of SEWe-focused FOBs to prioritize the well-being of the owning family over time (flow of SEW) may lead these firms to use a long-term horizon (Chrisman & Patel, 2012) in assessing the financial and socioemotional consequences of decline-stemming strategies (Abebe & Tangpong, 2018). As outlined in Table 1, decline stemming may adversely affect multiple stakeholder groups in the immediate term, thus inflicting a potential loss of SEW in the short term. However, evidence in the literature suggests that decline stemming can be considered an accommodating strategy when evaluating the impact on focal stakeholders over time (see Table 1, rows 5–7).
As Table 1 further shows, while decline-stemming strategies may negatively affect financial stakeholders in the short term, such strategies can be cash-saving or cash-generating over time. These may grant the firm much needed financial resources for recovery (Decker, 2016), thus allowing its long-term survival (Schmitt & Raisch, 2013). This desirable outcome benefits stakeholders (Table 1, rows 5 and 6) and makes transgenerational succession possible, a crucial component of SEWe. Furthermore, decline stemming can be a legitimacy-building tool (Abebe & Tangpong, 2018) and can stop the erosion of stakeholder support normally seen in turnaround situations (Arogyaswamy et al., 1995), thus representing an additional gain in SEWe. Therefore, we argue that stakeholders in an FOB that values SEWe are likely to react positively to decline-stemming strategies over time.
In summary, from an SEW mixed-gamble logic, we contend that significant SEWe benefits over time associated with the pursuit of decline-stemming strategies lead an owning family to weigh these gains more heavily than potential SEWe losses. Differently put, though a risky strategy, decline stemming is seen as a necessary strategy in increasing the flow of SEW (Chua et al., 2015). Gains in terms of flow of SEWe (future SEW) buffer the fear of diminished stock of SEW associated with such strategies (Chrisman & Patel, 2012). Importantly, these strategies may lead to a successful turnaround (Schmitt & Raisch, 2013), which would greatly improve the family’s financial resources and SEW. Conversely, failing to respond to a turnaround situation may trigger a downward spiral (Tangpong et al., 2015), with financially and emotionally devastating consequences for the owning family. We thus propose:
Method
Sample and Data Sources
We tested our hypotheses on a sample of family-controlled public firms traded on the main US stock markets (NASDAQ, NYSE, and AMEX). Consistent with previous studies, our sample excluded banks, financial institutions, and investment vehicles (e.g., closed-end funds, real estate investment companies, etc.) because their accounting measures are not comparable with those of nonfinancial firms due to the unique nature of their capital and investments.
We created the dataset in multiple stages. First, we relied on Capital IQ to screen for firms where an individual possessed at least 10% of the voting shares. Second, we retrieved accounting data from COMPUSTAT for a time window spanning 2000–2012 to encompass periods of both economic downturn and growth (Morrow, Johnson, & Busenitz, 2004). This resulted in an initial sample of 9,744 firm-observations (812 firms). Accounting data were then matched with data from Capital IQ, Hoover’s database, company proxy statements deposited at the SEC, and information from company websites. This procedure left us with 7,613 firm-observations with complete information across databases (635 firms). Third, following Pandit (2000), we screened data on a firm-by-firm basis to identify turnaround situations by using a 4-year cycle. Similar to previous studies (e.g., Morrow et al., 2004; Tangpong et al., 2015) a turnaround situation was detected if: (a) a firm had two consecutive years of return on investment (ROI) above the risk-free rate of return (RFRR), which prevented the inclusion of chronically underperforming firms, and (b) a firm had at least 3 years of ROI below the RFRR during the period of decline. RFRR is the threshold below which a firm is failing in economic terms as it is not providing a return for its shareholders. In addition, using a time frame of 3 years of consecutive decline increases the likelihood that the firm was truly in a turnaround situation as opposed to suffering a temporary downturn (Bruton, Ahlstrom, & Wan, 2003; Bruton, Oviatt, & White, 1994). Consistent with previous studies, our study employs ROI as a suitable measure of performance for turnaround situations (Bruton et al., 1994; Robbins & Pearce, 1992). Overall, these filtering procedures allowed us to identify 147 turnaround situations, or 147 four-year cycles corresponding to 588 firm-observations (a firm may have faced more than one turnaround situation over the 12-year period).
Following previous studies (Patel & Chrisman, 2014; Strike, Berrone, Sapp, & Congiu, 2015), we defined FOBs as firms in which members of the owning family held at least 10% of voting shares, at least one family member served as an insider, and the family had been involved with the firm for at least two generations. This classification of FOBs ensured an active involvement of family members in firm proceedings, thus enabling us to capture the owning family’s perceptions and beliefs using corporate narratives (Zachary, McKenny, Short, & Payne, 2011). The final sample thus consisted of 416 firm-observations (4-year data concerning 104 turnaround situations). We retained data from the onset of financial decline to the third consecutive year of decline.
Measures
We measured our dependent variables at year’s end for each year from the onset of performance decline until the end of the third year of that decline. The independent and control variables were collected creating a 1-year lag to ensure temporal ordering. The use of panel data helped us capture the trend in variation over time and provided a better understanding of the phenomenon (Barker & Mone, 1994; Robbins & Pearce, 1992).
Dependent variables
Cost retrenchment intensity (CRT)
This variable represented change in the firm’s cost base from 1 year to the next. The cost base consisted of sales minus cost of goods sold (CGS) minus operating income plus interest expenses (Barker & Mone, 1994; Morrow et al., 2004; Robbins & Pearce, 1992). We used the following formula to compute this variable:
where i = time (years 1–4), and j = firm.
Therefore, a positive value suggested cost retrenchment, and a negative value represented an increase in the firm’s costs.
Asset retrenchment intensity (ART)
Following previous studies (e.g., Morrow et al., 2004; Robbins & Pearce, 1992), we first calculated the asset base. This was defined as the sum of long-term and short-term assets (e.g., cash and equivalent, account receivables, inventory, and plant and equipment). We computed asset retrenchment as:
Therefore, a positive value suggested asset retrenchment, and a negative value indicated an increase in the firm’s assets.
Change in sales intensity (SCT)
This variable captured revenue generative strategies (Bruton et al., 2003). That is, strategies aimed at ameliorating the firm’s cash flow through increases in sales. Thus, we used change in sales from 1 year to the next expressed in percentage terms, which allowed us to avoid confounding effects due to firm size. Similar to previous studies (e.g., Bruton et al., 2003), change in sales was calculated as follows:
where i = time (years 1–4), and j = firm.
Thus, positive values suggested increases in sales and negative values showed decreases.
Independent variables: SEW construct validation
The measurement of multiple dimensions of SEW is a complex task. SEW is a latent construct deeply rooted in owning families’ value and belief systems (Berrone et al., 2012). Our research effort validates a measurement of SEW through CATA to overcome problems associated with its measurement (Short, Broberg, Cogliser, & Brigham, 2010) and provides support for aggregating the FIBER dimensions into two higher order constructs.
To extract the five SEW dimensions, we used CATA performed through DICTION, an increasingly popular software for this type of analysis in strategy and entrepreneurship studies (McKenny, Short, Zachary, & Payne, 2011; Short & Palmer, 2008). Specifically, we analyzed proxy statements and letters to shareholders deposited at the SEC. This method is widely used to capture the meaning behind the rhetoric employed in narratives (Short et al., 2010; Zachary et al., 2011). Evidence suggests that corporate narrative reflects the cognition of decision makers involved in the firm and can provide insights into management’s perception concerning the business (Short & Palmer, 2003; Zachary et al., 2011). Furthermore, previous family business studies have relied on the assumption that family involvement is highly correlated with the controlling family’s vision and objectives (Miller, Breton-Miller, & Lester, 2013; Patel & Chrisman, 2014; Strike et al., 2015) and that commitment to the business reflects a controlling family’s willingness to pursue family-centered behavior (De Massis, Kotlar, Chua, & Chrisman, 2014). Thus, a firm’s narrative (e.g., letters to investors, proxy statements, and annual reports) mirror mental models, beliefs, values, and perceptions of the firm’s dominant constituency (Brigham, Lumpkin, Payne, & Zachary, 2014; D’Aveni & MacMillan, 1990).
The use of CATA in our study is appropriate because we investigate publicly traded firms in which the family represents the dominant coalition. There is general agreement (e.g., Barr, Stimpert, & Huff, 1992; Brigham et al., 2014) that content analysis performed on letters to shareholders is able to capture the family’s values and perceptions (D’Aveni & MacMillan, 1990). Additionally, CATA relies on available texts, it is nonobtrusive, and it can be replicated (Berrone et al., 2012). Thus, it captures the complexity of the various SEW dimensions while avoiding most biases inherent in the use of survey methodology. Next, we outline the process of development and validation of our CATA measure for the SEW dimensions.
Content validity
We adapted Short et al.’s (2010) procedure to provide evidence for the validity and reliability of our measure of SEW. First, we defined content domain and dimensionality of this construct based on a literature review. We used Gómez-Mejía et al.’s (2007) definition of SEW and its FIBER conceptualization (Berrone et al., 2012). Following recent advances reported in the SEW literature (e.g., Debicki et al., 2016), we aggregated the FIBER categories into higher order SEW dimensions that could be reconciled with the two dimensions proposed by Miller and Le Breton-Miller (2014).
Second, as suggested in the literature (Brigham et al., 2014; Short et al., 2010), word lists (or dictionaries) representative of the various dimensions of SEW were generated. CATA uses word lists as a tool for measuring a construct. The lists were then evaluated by three scholars familiar with the SEW literature to assess the validity of each word and judge whether it captured the intended dimension of SEW. To gage interrater agreement, we calculated Holsti (1969) interrater proportion of agreement observed (PAo), which provided an assessment of the extent to which raters agreed that the words selected were appropriate to capture the underlying construct (Short et al., 2010). The PAo was calculated as follows: PAo = 3A/(n1 + n2 + n3), where A was the number of coding decisions on which the three raters agreed, while n1, n2, and n3 were the number of coding decisions made by the three raters, respectively. The interrater reliability coefficient was 0.9 overall (.82 for restricted SEW; .94 for extended SEW), thus suggesting that our dictionaries were appropriate for capturing the underlying construct (e.g., Brigham et al., 2014). As recommended by Brigham et al. (2014) and Short et al. (2010), we also engaged in an inductive process to detect additional words that would fit the various SEW dimensions. The Appendix (see Supplemental Material online) shows the final version of our word lists and descriptions of our dictionaries for the various dimensions of SEW. The word lists were then uploaded to the content analysis software DICTION 6.0 (Hart, 2000), which provided a score for each dimension of SEW. We standardized the scores to control for the number of words in the documents (e.g., Brigham et al., 2014; Zachary et al., 2011). This process led to more conservative estimates that were not biased by the length of the documents.
External validity
The validity of measures relying on CATA significantly depends on the narrative selected and the sampling procedure adopted (Brigham et al., 2014; Short et al., 2010). We consider our decision to rely on shareholder letters appropriate, for multiple reasons. First, a firm’s narratives mirror the mental models, beliefs, values, and perceptions of the firm’s dominant constituency (Berrone et al., 2012; D’Aveni & MacMillan, 1990), much like SEW, which is rooted at a deep psychological level in family owners’ minds (Berrone et al., 2012). In addition, the strategic behavior of publicly traded family firms in which the owning family represents the dominant coalition (Faccio & Lang, 2002) is largely influenced by the preferences, experiences, and desires of the family owners. Second, shareholder letters are widely used in the literature to measure latent constructs in FOBs (e.g., Brigham et al., 2014; McKenny et al., 2011). As D’Aveni and MacMillan (1990) explain “…letters to shareholders are particularly good indicators of the major topics that organizational managers attend to” (p. 640). Third, our method enabled us to capture the complexity of the various SEW dimensions (Berrone et al., 2012). Last, we acknowledge that shareholder letters may to some extent be influenced by impression management techniques and public relations strategies. Nevertheless, empirical evidence (e.g., Amernic, Craig, & Tourish, 2007) suggests that top executives within an organization actively participate in the letter writing process. For this reason, we used family involvement as a criterion in selecting the FOBs in our sample. Furthermore, even when the authorship of such documents is not specified, the dominant coalitions within the firm significantly shape their content (Barr et al., 1992; Brigham et al., 2014). Thus, our analysis is expected to capture the owning family’s values and perceptions (D’Aveni & MacMillan, 1990).
In addition, we created a validation sample of 60 FOBs drawn from the “Top 100 Family Businesses in North America” and from the US firms found in the “Top 250 Multinational Family Firms” list (Casillas & Pastor, 2015), and we content-analyzed letters to shareholders relying on our custom dictionaries. The values returned by DICTION were used to perform empirical tests. First, a one sample t-test performed for each dimension of SEW found that all five dimensions of SEW were significant (Table 2), which suggested that shareholder letters across samples included language consistent with SEW. Second, a one-way analysis of variance (ANOVA) test (Table 3) suggested no significant differences in SEW dimensions across the two samples of family businesses. The exception was binding social ties, which was only marginally significant.
Evidence of Language Representing SEW in Corporate Narratives of FOBs in Main Sample and in the Top Multinational Family Firms Database.
Note. The dimensions of SEW in this table were captured using CATA methodology on corporate narrative relying on the dictionaries explained in the Appendix (see Supplemental Material online). FOB = family-owned business; SEW = socioemotional wealth; SD = standard deviation.
**p < .001.
ANOVA Comparison of Firms From Main and Validation Samples on SEW Dimensions.
Note. The dimensions of SEW in this table were captured using CATA on corporate narratives relying on the dictionaries explained in the Appendix (see Supplemental Material online). CATA = computer-aided text analysis; SEW= socioemotional wealth; ANOVA = analysis of variance.
aTo rule out the possibility that differences between the two samples may be influenced by the different sample sizes, we created two random subsamples of 60 firms.
*p < .1
Dimensionality. Dimensionality refers to relatedness among multiple dimensions of a single construct and their association with the unified construct. Following Short et al. (2010), we assessed dimensionality by using a correlation matrix and factor analysis of the DICTION scores. As shown in Table 4a, all dimensions of SEW (except Family Control and Influence) present a nearly perfect intercorrelation (i.e., higher than .5, and significant). This suggests that SEW comprises two distinct dimensions: restricted (SEWr) and extended SEW (SEWe). SEWe encompasses four of the five dimensions of SEW (i.e., IBER). This conceptualization was also supported by a factor analysis (see Table 4b). The result suggests that the IBER dimensions of SEW have low uniqueness and load on the same factor, supporting their aggregation into a higher order dimension (Brigham et al., 2014). Internal consistency among items for SEWe was satisfactory (α = .86). In contrast, F loaded on a distinct factor and had high uniqueness; thus, it should be considered a distinct dimension.
(a) Correlation of SEW Dimensions Performed to Assess Dimensionality and (b) Factor Analysis for SEW to Test Dimensionality (Family Firms).
Note. SEW = socioemotional wealth.
**p < .001.
It is worth noting that although we found support for two dimensions of SEW, emotional attachment was surprisingly associated with SEWr rather than SEWe. We ascribe this to the fact that E captures a broad spectrum of affective benefits, some of which are conceptualized as elements of SEWr and others as part of SEWe. We suggest that elements of E, such as the bond that exists among family members and accounts for stable relationships, or the ties that help maintain a positive self-concept (e.g., Hauck et al., 2016) require the firm to be focused on long-term considerations (Chrisman & Patel, 2012). Thus, the E dimension of SEW encompasses both flow and stock characteristics (Chua et al., 2015). Our operationalization of E placed greater emphasis on the affective value of E over time (flow of SEW) and hence captured elements closer to the priorities of an owning family focused on R and I. For this reason, we believe that the limitations of our measure of E notwithstanding our results were not biased. However, we suggest that future research may draw from our findings to propose a conceptualization of the E dimension of SEW capable of separating its flow and stock components. Doing so would allow for the attribution of stock and flow elements of E to SEWr and SEWe, respectively, thus fully reconciling the FIBER dimensions and our conceptualization of SEW.
Concurrent validity
Following previous scholars (e.g., Brigham et al., 2014), we provide evidence for concurrent validity by comparing our measurement of SEW to family business status, a variable expected to appear concurrently with SEW. To this end, we created a matched sample of family and nonfamily firms and tested differences in the level of SEW with a between-groups ANOVA. The two groups showed significant differences in SEW rhetoric (Table 5). These results were consistent with Berrone et al.’s (2012) argument and thus supported the validity of our measures for the SEW dimensions.
One-Way ANOVA for Between-Group Differences in Levels of SEW of Family and Nonfamily Firms.
Note. This table uses the conceptualization of SEW with two dimensions, as supported in Table 4. ANOVA = analysis of variance; SEW = socioemotional wealth; MS = mean square.
Controls
The data analysis controlled for variables suggested by previous studies as being influential on our dependent variables. It was critical to control for industry membership, firm size, internally generated decline, and financial slack because those variables have a significant impact on the type and intensity of turnaround strategies (e.g., Bruton et al., 2003; Pearce & Robbins, 1993). Thus, we included the following control variables: a firm’s SIC dummy-coded at the second-digit level (nine categories), the natural logarithm of sales, a dummy variable coded one if firm performance was above the industry average (Robbins & Pearce, 1992), and current ratio. Evidence suggests that stakeholders are more prone to providing their support to firms they perceive as older and historic (Choi & Shepherd, 2005). Thus, our models controlled for the difference between the year of onset of financial decline and the firm’s foundation year. Diversification and international diversification also affect a FOB’s perception of risk, both systematic and unsystematic (Gómez-Mejía, Makri, & Kintana, 2010). Hence, the regression controlled for entropy as well as the volume of foreign sales expressed as a percentage of total sales (Gómez-Mejía et al., 2010). We also controlled for variables deemed to limit a family’s ability to impose its unconstrained will on the firm (e.g., Strike et al., 2015) or alter a family firm owner’s perception of the business. Thus, we included in our analysis institutional shareholding (Lai & Sudarsanam, 1997) as well as a dummy variable coded one if the firm had the family owner’s name (Gómez-Mejía et al., 2010) and 0 otherwise. Lastly, we controlled for time effect with a series of time dummies (four categories for time 1–4).
Data Analysis
To overcome the methodological challenges resulting from the use of panel data, we ran multiple models relying on the feasible generalized least squares (FGLS) approach (Wooldridge, 2012). Our FGLS models were specified to correct for the potential presence of panel heteroscedasticity and panel common serial correlation. The descriptive statistics and the correlation coefficients for the variables in our model are shown in Table 6.
Descriptive Statistics and Correlations (N = 416).
Note. The total sample encompasses 416 firm-observations (104 turnaround situations); 238 firm-observations concerned firms that engaged in asset retrenchment while 178 concerned firms that did not, 235 firms that engaged in cost retrenchment while 181 concerned firms that did not), and 265 engaged in sales increases while 151 concerned firms that did not). Note also that there was significant variation in the level of assets, costs, and sales changes. SD = standard deviation; SEW = socioemotional wealth.
*p < .05. **p < .01.
Model estimation and results
We tested Hypothesis 1 (H1) for the effects of SEWr on the intensity of individual decline-stemming strategies while controlling for SEWe (Table 7). As we show in Model 1, the coefficient for our indicator of SEWr was negative (β SEWr = −382.75) and significant (p < .001); Model 1 was significant overall (Wald χ2 = 1033.98, p < .001). This supports a negative relationship between SEWr and asset retrenchment (H1a). Similarly, Model 2 showed a negative and significant relationship (β SEWr = −241.78; p < .01) between SEWr and cost retrenchment (H1b), and the model was overall significant (Wald χ2 = 14252.90, p < .001). Therefore, H1b was supported. Last, our test supported a negative relationship between SEWr and sales increase. The coefficient for SEWr (Model 3) was negative and significant (β SEWr = −609.44; p < .001), and Model 3 was significant overall (Wald χ2 = 1153.30, p < .001).
Feasible Generalized Least Squares a Models for Decline-Stemming Strategies in FOBs.
Note. (a) The model corrects for autocorrelation and heteroskedasticity of the residuals. (b) All variables were lagged except for those indicated with a superscript b. (c) The inverse Mills ratio was calculated using Heckman selection models; the selection equation included: lagged family ownership, lagged family directors, lagged family to independent directors ratio, lagged generational involvement, family CEO duality dummy, family CEO dummy, family name dummy, industry effect (nine categories), time effect (four category), internally generated decline (dummy), lagged Tobin’s Q, diversification, international diversification, family firmtype (four categories), lagged institutional ownership, founder-centered family firms (dummy), firm age, firm size, and lagged financial slack. (d) There was a 1-year lag between independent and dependent variables; therefore, the number of firm-observation is 312 (i.e., 416 firm-observations minus 104 cases).
*p < .05. **p < .01.
We followed the same procedure to test for Hypothesis 2 (H2), which concerned a positive relationship between SEWe and intensity of decline-stemming strategies. As shown in Table 7, for Model 1 predicting asset retrenchment (β SEWe = 6.90, p < .01), Model 2 predicting cost retrenchment (β SEWe = 15.10, p < .001), and Model 3 predicting sales increase (β SEWe = 30.583, p < .001), the coefficients for our indicator of SEWe were positive and significant while controlling for SEWr. Furthermore, all three models were significant overall. Therefore, H2 (a, b, and c) is supported.
In conclusion, our hypotheses testing provides interesting insights. Family firms that scored high on our measure of SEWr tended to display highly family-centric behaviors in the form of reluctance toward the pursuit of intense decline-stemming strategies. In contrast, family firms that scored high on our indicators of SEWe showed a willingness to nurture the firm (Miller & Le Breton-Miller, 2014) by engaging in significant asset and cost retrenchment in tandem with sales increases with the aim of turning the firm around.
Robustness checks
We checked for the potential presence of biases (Block, 2010). Our sample comprised firms in need of turnaround regardless of whether they engaged in a strategic response to such decline in performance. In such situations, the use of a Heckman (1979) model represents the ideal solution for countering selection bias. The results of our robustness checks suggested that our analyses were not substantially threatened either by sample selection or by reverse causality bias. In addition, we ruled out possible alternative explanations by running our main analysis afresh with the addition of objective measures of family control over the business (i.e., the percentage of voting shares held by the family, family involvement, and the ratio of family members to independent directors). Similarly, we controlled for the possibility that a firm’s debt load (measured as debt-to-equity ratio or long-term leverage) may increase financial stakeholders’ power to steer the firm’s turnaround response (e.g., Jawahar & Mclaughlin, 2001; Pajunen, 2006). The addition of these variables did not alter our results.
Discussion and Conclusion
Our study sought to examine the impact of nonfinancial benefits (SEW) on the strategic behavior of family firms in need of turnaround. The investigation of a sample of 416 turnaround situations indicated that FOBs appear to frame their decisions to preserve the support of those focal stakeholders who helped the owning family to continue deriving SEW. Specifically, an owning family that valued SEWr ascribed importance to family stakeholders in order to preserve the short-term financial and socioemotional interests of the immediate family. In the context of decline-stemming strategies, SEWr preservation indicates a preference for low-risk strategies that accommodate the claims of financial resource providers (Decker, 2016) and prevent losses of SEWr, which are perceived to outweigh potential gains. Our findings are consistent with previous studies, which reported a negative relationship between family business status and turnaround strategies such as divestment and down-scoping (Chung & Luo, 2008), downsizing (Block, 2010), and wage cuts (Essen et al., 2015). These studies posited that the owning family’s sensitivity toward certain stakeholders’ claims (e.g., employees) led to a reluctance to engage in turnaround and increased the risk of bankruptcy.
Our findings are also in line with the literature on family firm resilience (Chirico et al., 2018) during drastic performance decline due to a willingness to engage in significant restructuring (Kavadis & Castañer, 2015), pursue recovery strategies, and demonstrate commitment toward the business (Chirico et al., 2018). We found that family firms that valued SEWe pursued decline-stemming strategies that protected the interests of the family and critical stakeholders in the long term. It appears that the prospects of long-term gains in SEWe outweigh fears of diminished SEWe in the immediate term. Therefore, our findings go a long way toward reconciling conflicting positions on FOBs turnaround. Our study is an important first step in the process of enhancing our understanding of the role played by nonfinancial objectives in FOBs’ strategic decision making in response to the demands of different stakeholder groups. Thus, our study makes a meaningful contribution for both theory and practice.
Implications for Theory and Practice
Our first contribution entails a robust theoretical framework for explaining heterogeneity in FOBs’ strategic behavior. Our argument showed that stakeholder theory can be informed by insights from the SEW literature to explain the strategic preferences of FOBs while taking the multidimensionality of SEW into consideration. Although this is an incremental theoretical contribution (Reay & Whetten, 2011), we enhance the explanatory power of current theoretical frameworks, which may help scholars in the investigation of other FOB phenomena.
By investigating the decline-stemming strategies in FOBs, we extend Jawahar and Mclaughlin’s (2001) argument that firms under conditions of decline simply look at who among the stakeholders provides economic benefits to the firm and ignore others. In fact, we suggest that FOBs undertake a specific turnaround strategy because of the noneconomic benefits particular stakeholder groups can provide. This perspective is in line with recent studies, which proposed that family firms strive to enhance or protect the accumulated stock of SEW (Cruz & Justo, 2017). In addition, since SEW is a multifaceted construct, the identification of the most relevant stakeholders is contingent upon which type of SEW is used as a frame of reference by the family principal.
Our second contribution stems from the multidimensional conceptualization of SEW and from the validation of its measurement. To date, scholars have not coalesced around a common definition and operationalization of SEW (Berrone et al., 2012). Clearly, this inhibits the advancement of family businesses as a field of study, which requires scholars to improve definitions and measurements of its core constructs (Reay & Whetten, 2011). In response, our research adapted and incorporated a recent conceptualization, which offers a parsimonious categorization of SEW into SEWr and SEWe along with previous approaches (e.g., the FIBER model). This effort goes a long way toward helping scholars strengthen family business theory and capturing the complexity of the SEW construct, which may help theory building through better explanatory power and promote the reconciliation of mixed findings.
Our investigation makes other important contributions. We heeded a recent call to engage in within-family-firm comparisons and to strive to capture heterogeneity within this organizational form (Salvato & Aldrich, 2012). Additionally, our study advances previous investigations that attempted to disentangle the multidimensional nature of SEW and the strategic preference of FOBs based upon the affected stakeholder groups (e.g., Cruz et al., 2014; Vardaman & Gondo, 2014). Unlike earlier studies, we identify focal stakeholders by measuring the value a specific owning family ascribes to certain nonfinancial benefits (i.e., SEWe, SEWr). Thus, our study confirms Vardaman and Gondo’s (2014) conjecture that FOBs need to balance the tension between financial and nonfinancial (SEW) objectives as well as between different types of SEW. Since the values ascribed to SEWe and SEWr vary from one family firm to the other, we infer that the type of SEW deemed more important will impact the frame of reference FOBs use in decision making (Debicki et al., 2016). Hence, there appears to be no general pattern for managing SEW conflict (Vardaman & Gondo, 2014). Future research may delve deeper into SEW conflicts in FOBs by investigating conditions under which the combination of value ascribed to SEWr and SEWe changes.
Our study also informs practice. First, we provide insights regarding how prioritizing nonfinancial objectives could drive a firm’s strategic behavior when family wealth is at risk. Recognizing that different family principals operating under different frames of reference could bolster a specific strategy may help them better anticipate and address within-family conflicts typical of family firms in crisis situations. Thus, family owners may use our findings to avoid mistakes that could exacerbate the firm’s decline. Second, our approach may be useful to nonfamily stakeholders (e.g., minority shareholders) as family principals may decide to let the claims of less important stakeholders go unaddressed. Our model could provide a firm’s stakeholders with the ability to predict the strategic behavior most likely to be undertaken based on the family firm’s main frame of reference.
Limitations and Future Research
Our work presents some limitations. Although we were unable to provide support for the five dimensions of SEW proposed in the FIBER model (Berrone et al., 2012), we showed that the FIBER dimensions can be reconciled with the two higher-order SEW constructs proposed by Miller and Le Breton-Miller (2014). This results in a more parsimonious typology of SEW. Additionally, rather than relying on primary data, we measured SEW by using CATA on corporate narratives from publicly traded firms (Debicki et al., 2016; Hsueh, 2016). Our approach is a widely accepted alternative for capturing beliefs, values, and perceptions of the firm’s dominant constituency (Brigham et al., 2014) in a nonintrusive way. This is because the collection of primary data is likely to hurt the validity of the responses as well as the response rate (Zachary et al., 2011). Future research efforts may use available psychometric measures (Berrone et al., 2012; Debicki et al., 2016) to capture all FIBER dimensions and attempt to connect them to FOB responses in turnaround situations. Scholars may also evaluate whether a less parsimonious categorization of SEW is justified by the ability to better understand the strategic preferences of FOBs.
The approach and theoretical model adopted in our study may be applied to explain a broad array of family business phenomena. For instance, scholars may use our theoretical model to investigate the second stage of turnaround response (i.e., recovery). Specifically, our framework may help to identify critical stakeholders for SEWr- and SEWe-oriented FOBs during the recovery stage of turnaround. In turn, this may help scholars better explain and predict the type and intensity of recovery strategies being pursued while also considering other important contingencies (e.g., competitive position and root causes of the decline). More generally, our approach is not idiosyncratic to turnaround in family firms since it may be applied to explain the heterogeneity of family-firm behavior in all those situations where the fulfillment of claims from a particular group of stakeholders may affect SEWr and SEWe differently (e.g., environmental practices, socially responsible practices, philanthropic activities, etc.). It would also be interesting to explore the performance consequences of decline-stemming strategies as they are influenced by the pursuit of either SEWr or SEWe. Last, delving into conflict between different types of SEW was beyond the scope of this study. Future research may explore whether—and if so, how—the importance ascribed to different types of SEW changes over time. Further, the understanding of how family firms address SEW conflict when equal importance is ascribed to different types of SEW may represent a fruitful avenue for future research. We believe that our study offers the theoretical and empirical tools needed in the exploration of these research avenues.
Supplemental Material
Appendix - Supplemental material for The Impact of Socioemotional Wealth on Decline-Stemming Strategies of Family Firms
Supplemental material, Appendix, for The Impact of Socioemotional Wealth on Decline-Stemming Strategies of Family Firms by Giacomo Laffranchini, John S. Hadjimarcou, and Si Hyun Kim in Entrepreneurship Theory and Practice
Footnotes
Acknowledgments
The authors gratefully thank the two anonymous reviewers and the Editor Donald Neubaum for their constructive comments and insightful suggestions that helped strengthen this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
