Stakeholder engagement strategies, national institutions, and firm performance: A configurational perspective

**Research summary** 
Research documents the performance effects of attending to shareholders and treating employees well but underplays national differences in the relative power of labor and capital. We advance a configurational perspective that acknowledges the fit between stakeholder engagement, context, firm attributes and performance. As a cornerstone of this perspective, we develop a typology of stakeholder engagement strategies expressing how firms navigate the tension between conforming with local expectations—by prioritizing shareholders or employees, according to context—and being distinctive—by diverging from their peers. Analyzing a cross‐national sample of firms from 2004 to 2011, we identify combinations of engagement strategies, firm attributes, and contexts linked to high performance. Our findings highlight the multiple context‐dependent paths, which link stakeholder engagement to high firm performance. 
 
**Managerial summary** 
How do firms navigate pressures from shareholders and employees across different institutional environments? We develop a typology of stakeholder engagement strategies based on how firms in different countries strike a balance between conformity (i.e., prioritizing locally important stakeholders) and differentiation (i.e., prioritizing stakeholders that their local peers might neglect). Our findings show that the engagement strategies associated with high performance vary according to local institutional context and firm characteristics. In particular, by not merely prioritizing stakeholders who are already locally important, firms can use stakeholder engagement to differentiate themselves from their peers, and such engagement strategies are often linked to high performance. 
 
**Online appendix: Data Set** 
Available at [https://doi.org/10.35065/sten-2001](https://doi.org/10.35065/sten-2001)


| INTRODUCTION
Institutions-"the rules of the game in a society or….the humanly devised constraints that shape human interaction" (North, 1990, p. 3)-both constrain and empower actors and activities (Scott, 2001). Institutions linked to the stakeholder groups of labor and capital are of high importance for firms (Aguilera & Jackson, 2003;Ahmadjian & Robbins, 2005), and the relative importance of these two stakeholder groups is the core distinction between market-driven and coordinated institutional systems. In market-driven institutional systems, firms are reliant on equity financing and have access to flexible labor markets, and the state enshrines firms' fiduciary duty to their shareholders in law. In coordinated institutional systems, firms rely on longterm bank-based financing, provide greater employment security for their employees, and the state restricts employers' rights to unilaterally change employment terms. These cross-national differences in labor relations and the working of capital markets shape firms' relationships with their shareholders and employees (Coates, 2005;Hall & Soskice, 2001;Witt & Jackson, 2016).
It is common for competing interests to produce tensions between labor and capital (Coff, 1999;Falato & Liang, 2016). Market-driven and coordinated institutional systems represent institutional attempts to address these conflicts, but within both types of environments firms retain some discretion in how they allocate resources among stakeholders (Mitchell, Agle, & Wood, 1997). For example, firms might use employment quality initiatives to attract talent (Bode, Singh, & Rogan, 2015) or governance initiatives to raise new capital (Gompers, Ishii, & Metrick, 2003). Firms with limited resources must typically choose between such initiatives, and performance outcomes are likely to be sensitive to the institutional context in which a firm operates (Barnett, 2007). Prior research has assessed the consequences of employee treatment (Edmans, 2011) or governance reform (Tuschke & Sanders, 2003) on firm performance. Although this research has provided important insights, it has generally studied each aspect in isolation (e.g., initiatives for labor or capital) and in a single institutional setting (e.g., the United States or Germany). Therefore, the existing literature does not enable insights into how firms across different contexts simultaneously address pressures from labor and capital, and how doing so impacts their performance.
Addressing this gap in the literature, our study's research question is as follows: under which institutional and firm-level conditions are different stakeholder engagement strategies linked to high financial performance? Our point of departure is the importance of fit between strategy and context. We develop a typology of engagement strategies that combines two theoretical premises: (a) that institutional contexts generate rules and norms that shape stakeholder engagement and (b) that these norms create opportunities for distinct strategies. The combination of these two premises is vital because firms perform well when they are sufficiently different from competitors while also conforming to local expectations (e.g., Deephouse, 1999;Phillips & Zuckerman, 2001;Zhao, Fisher, Lounsbury, & Miller, 2017). Accordingly, our typology holds that firms can focus their stakeholder engagement: (a) on activities that complement local institutional arrangements ("complementary engagement"), (b) on stakeholders that are of lesser importance locally ("substitutionary engagement"), (c) on adhering to minimal institutional requirements ("minimalist engagement"), or (d) on activities that both complement and substitute for their institutional environments ("encompassing engagement"). Building on this typology, we develop a configurational model that recognizes that firm's stakeholder engagement strategies are linked to performance in combination with firm-level factors, such as the degree to which maintaining the firm's business is dependent on support from employees and from shareholders.
To study this configurational model, we use fuzzy-set qualitative comparative analysis (fsQCA; Ragin, 2000Ragin, , 2008. This approach enables us to assess how distinct stakeholder engagement strategies are linked to firm performance under different firm-level and institutional conditions. With a data set of 122 firms across 13 countries for the period from 2004 to 2011, we identify five configurations of stakeholder engagement and institutional context that are consistently linked with high performance. These configurations capture different conditions in which substitutionary, minimalist, or encompassing engagement strategies are linked to high performance, whereas no configuration linked to high performance involves a complementary engagement strategy. Accordingly, our central contribution is to demonstrate that, beyond the normative stakeholder management defined by national institutions, firms may pursue one of a number of engagement strategies. Our findings highlight a key role for strategy, contingent on institutional context and firm-level factors, in linking stakeholder engagement to financial performance. Scholars have called for researchers to take into account environmental complexity in order to understand how firms manage the tension between conformity and differentiation (Zhao et al., 2017). By assessing a range of environments, we find that, rather than pursuing a single "optimally distinct" strategy that strikes a balance between conformity and differentiation, firms can pursue other engagement strategies that fit their particular context.

| THEORETICAL BACKGROUND
The management of stakeholders-groups or individuals "who can affect or [are] affected by the achievement of the organization's objectives" (Freeman, 1984, p. 46)-is a crucial element of strategy. Research on the relationship between stakeholder management and firm performance (Margolis, Elfenbein, & Walsh, 2007;McWilliams & Siegel, 2001;Orlitzky, Schmidt, & Rynes, 2003) is inconclusive, which suggests that it is neither simple nor universal. Accordingly, we build on recent research emphasizing firm performance as the outcome of a complex, multilevel process (Fiss, 2007(Fiss, , 2011Misangyi et al., 2017), dependent on both firm strategy and its institutional environment. The resulting configurational model considers the joint effects of country-and firm-level dimensions to explore the implications of how firms attend to stakeholder groups representing labor and capital. We focus on employees and shareholders as the core factors of production (Cobb & Douglas, 1928) and key actors in the institutional environment (Aguilera & Jackson, 2003). Our model is depicted in Figure 1 and described in detail below. We first discuss differences in institutional contexts as the backdrop for firm action. We then outline a typology of stakeholder engagement strategies that firms can pursue. Finally, we discuss firm-level factors that promote a focus on specific types of stakeholders.

| National institutional contexts and stakeholder engagement strategies
The varieties of capitalism literature synthesizes institutional differences across countries into two ideal types: liberal market economies (LMEs) and coordinated market economies (CMEs) (Hall & Soskice, 2001;Judge, Fainschmidt, & Brown, 2014). In LMEs, firms rely on market-based arrangements for coordination and financing, the fiduciary duty toward shareholders has legal and normative force, and interfirm relationships are characterized by arm's-length exchanges. The United States and the United Kingdom are the countries closest to this ideal type. In CMEs, banks are an important source of financing, and shareholders are only one of a variety of stakeholders protected by corporate law, which grants employees particular importance (Jürgens, Naumann, & Rupp, 2000). Firms in CMEs use extensive nonmarket mechanisms for coordination and rely more on relational approaches than firms in LMEs. Germany, Austria, and Norway are the countries closest to this ideal type, although market-based coordination has become increasingly important in several CMEs, including Germany and Denmark (Hall & Thelen, 2009;Rose & Mejer, 2003).
The differences between capitalist regimes have vital implications for stakeholder engagement (Kang & Moon, 2011;Matten & Moon, 2008) and its link to performance. First, institutional contexts generate rules and norms that influence which stakeholders are accorded importance (Ioannou & Serafeim, 2012;Maignan & Ralston, 2002). In LMEs, shareholders are supported more strongly than employees, whereas the opposite is found in CMEs. Second, by shaping stakeholder expectations, the rules and norms prevalent in a particular context create distinct opportunities that affect the financial outcomes of a given stakeholder engagement strategy. However, although national institutions and their dominant norms influence organizational action and the reactions of stakeholders (Brammer, Jackson, & Matten, 2012), firms have discretion in the choices they make regarding their stakeholder relationships (Barnett & Salomon, 2012;Tantalo & Priem, 2016).
The dashed lines between institutionlevel and firm-level drivers of engagement and firm performance represent direct (as opposed to configurational) relationships that are not the focus of our study [Color figure can be viewed at wileyonlinelibrary.com] The tension between firms' needs to comply with their institutional environment and to distinguish themselves presents firms with a choice to complement or substitute for the institutional context when designing stakeholder engagement strategies. For instance, the Norwegian context provides strong institutional support for employees, reflected in norms around long-term hiring and workers' board representation. Shareholders are only one of several important stakeholder groups. Thus, Norwegian firms may pursue complementary stakeholder engagement by focusing on serving the interests of locally dominant stakeholders, that is, employees. Alternatively, they may pursue substitutionary stakeholder engagement by focusing on the interests of stakeholders of relatively lower importance in this institutional context, that is, shareholders.
Based on these two strategic options, two additional stakeholder engagement strategies can be identified. First, firms may pursue a minimalist approach by not undertaking substantive stakeholder engagement activities for any stakeholders. For Norwegian firms, this would entail not undertaking additional activities to serve either shareholders or labor. Second, firms may pursue an encompassing approach by focusing on both the emphasized and less-supported stakeholders in their institutional context. For Norwegian firms, this would entail undertaking activities targeted toward both shareholders and labor. These four strategies are summarized in Table 1 and are elaborated on below.

| Complementary stakeholder engagement
Firms may exploit complementarities between their activities and dominant institutional forces. For example, in the United States, a country with strong capital institutions, firms may gain superior access to capital and consequently improve their performance by paying above-average attention to shareholder rights (Gompers et al., 2003). However, evidence suggests that even many high-performing firms place some restrictions on shareholder power (Gompers et al., 2003). Whereas practices conforming to institutional norms promote organizational survival (Zucker, 1987), they do not as a rule produce superior financial returns. Firm-level factors, such as local visibility (Phillips & Zuckerman, 2001), act as boundary conditions.

| Substitutionary stakeholder engagement
Firms may alternatively direct their engagement to stakeholders that are relatively overlooked by local institutions (Jackson & Apostolakou, 2010). For example, in contexts featuring weak institutional protection of labor, activities tailored to employee welfare enable firms to differentiate themselves from their competitors. Thus, firms may benefit from improving working conditions even when doing so is not locally mandated (Edmans, Li, & Zhang, 2014). Conversely, some firms in countries where institutions have historically provided stronger support to labor than shareholders implement stronger governance and protection of shareholder rights (Ahmadjian & Robbins, 2005;Fiss & Zajac, 2004) in the hope of raising capital.

| Minimalist stakeholder engagement
Firms following a minimalist approach do not undertake stakeholder engagement activities beyond those that are legally mandated. For example, in hybrid institutional contexts that combine norms supporting shareholder rights with strong employee rights (Kogut & Ragin, 2006), firms may decide to forego the diminishing marginal benefit of investing additional resources to attend to the interests of either shareholders or labor (for similar arguments, see Garcia-Castro &Francoeur, 2016, andHusted &de Jesus Salazar, 2006).

| Encompassing stakeholder engagement
An encompassing stakeholder-engagement approach simultaneously focuses on activities that complement and substitute for firms' institutional contexts; for example, by undertaking activities to serve both shareholders and labor. Indeed, stakeholder advocates originally called for attentiveness to a broad range of stakeholders (Freeman, 1984). This strategy might be especially beneficial for firms exposed to high levels of scrutiny. For instance, firms that primarily rely on their domestic market may face high expectations for conformity but may also benefit from attending to stakeholders who are less-supported in their institutional contexts. These four strategies reflect different approaches to achieving either relative conformity or relative differentiation with respect to firms' institutional environments. A complementary strategy denotes an approach to achieve relatively higher conformity by focusing on engaging stakeholders supported by local institutions while not focusing on stakeholders neglected in that institutional context. Conversely, a substitutionary strategy is an approach that emphasizes achieving relatively higher differentiation by taking initiatives to focus on stakeholders not supported by local institutions and by not focusing on stakeholders that are supported by them. An encompassing strategy, focusing on both shareholders and employees, balances conformity with differentiation by focusing on both stakeholders relatively neglected and supported by local institutions. A minimalist strategy, focusing on neither shareholders nor employees, strikes yet another kind of balance, by neglecting institutionally powerful stakeholders while not devoting additional resources to institutionally neglected stakeholders. The latter two instances correspond to compensatory orchestration (Zhao et al., 2017), which suggests that a firm becomes "optimally distinct" by conforming on one dimension while differentiating on another.

| Strategic drivers of stakeholder engagement
In addition to the opportunities for conformity and differentiation captured by our typology, the outcome of stakeholder engagement activities also depends on firm-level drivers, such as the degree to which a firm is dependent on support from employees and shareholders. We integrate three firm-level factors into our configurational model. These relate to pressure to attend to employee and shareholder concerns, i.e., the need for specialized labor and the firm's ownership concentration; and the extent to which the firm is embedded in its domestic institutional system (i.e., internationalization). In selecting this parsimonious set of firm-level factors, we also consider the empirical limits to the number of attributes to include in a research design utilizing the set theoretic approach of qualitative comparative analysis (QCA) (Marx & Dusa, 2011).

| Need for specialized labor
Incentives for firms to address employee interests are shaped by their core activities (Datta, Guthrie, & Wright, 2005), which influence the extent to which they can rely on external labor markets and generalist training systems. Firms engaging in high-technology activities require specialized labor, and their focus on R&D characterizes the kind of employment they provide (Keeble, 1990). Accordingly, high-tech firms stand to gain more from the increased retention and productivity that results from superior employee treatment (Edmans et al., 2014;Miller, Lee, Chang, & Le Breton-Miller, 2009). Employees are likely to reciprocate with additional effort, knowledge, and/ or inputs if they perceive that a firm acts fairly toward them (Bosse, Phillips, & Harrison, 2009).
Thus, for firms requiring specialist labor a focus on employee engagement may represent a viable strategy. However, this stakeholder engagement strategy is not likely to provide the same benefits for firms engaged in low-technology activities. For these firms, hiring low-skilled or generalist employees and managing the mobility of this workforce is less challenging, and the recruitment pool is wider. Where labor institutions are already strong, lower mobility and the comparatively better conditions already offered to employees reduce the benefits that a firm can derive from differentiating by means of employee initiatives. However, even in these contexts, firms that have a higher need for specialized labor will gain more from attention to employees than firms that are engaged in low-tech activities.

| Concentrated ownership
Not all firms need to be equally responsive to shareholders' interests. Ownership structure affects not only a firm's performance (Thomsen & Pedersen, 2000), but also how it engages with stakeholders, and especially its shareholders (Stavrou, Kassinis, & Filotheou, 2007). Because the interests of controlling and minority shareholders are often misaligned (Enrique & Volpin, 2007), whether a firm has a few large owners or widely dispersed ownership affects its incentive to improve minority shareholder rights (la Porta, Lopez-de-Silanes, Shleifer, & Vishny, 2000).
Firms with a large controlling ownership generally have less incentive than other firms to focus on the interests of the wider body of shareholders or to pursue governance reform. By curtailing the rights of minority shareholders, a few large shareholders can establish a high degree of effective control, allowing them also to effectively expropriate minority shareholders (la Porta, Lopez-de-Silanes, & Shleifer, 1999). In addition, higher control by few shareholders allows more effective monitoring of managers (Pagano & Röell, 1998). However, the outcome of ownership concentration should be interdependent on firms' institutional environments: a lack of shareholder protection provided by a country's institutional environment may present firms with more opportunities to differentiate themselves and gain superior access to financing (Gompers et al., 2003).

| Internationalization
Many large firms eventually expand their activities beyond their home countries. Internationalized firms face differing pressures across distinct institutional contexts, which may affect which stakeholders they attend to (Devinney, 2009;Durand & Jacqueminet, 2015) and the performance outcomes of stakeholder engagement approaches (Campbell, Eden, & Miller, 2012). Multinationals have greater discretion than firms that only operate domestically because the former can more readily relocate activities when conforming to local norms is costly (Scherer, Rasche, Palazzo, & Spicer, 2016). A priori, firms that rely on international markets face less pressure to conform locally because a global audience is unlikely to care whether they meet standards set by their respective home country's institutional context. However, highly internationalized firms also face overseas scrutiny, sometimes leading to engagement activities for multiple stakeholders with the aim of being responsive across contexts (Crilly, 2011).

| METHODS AND DATA
Our research design considers the interdependencies between country-level institutional and firm-level factors by exploring how their combined impact is linked to firm performance. Accordingly, we adopt fsQCA (Ragin, 2000(Ragin, , 2008 to identify configurations of stakeholder engagement activities, national institutions, and firm-level factors that are linked to high performance. This approach recognizes the potential for causal complexity underlying organizational phenomena, including how multiple factors are together linked to an outcome (in the present study, performance) and that there may be multiple pathways linked to a particular outcome (Misangyi et al., 2017).

| Sample and data
To construct our sample, we identified the 100 largest firms by market capitalization in 2012 in each of 13 countries: Australia, Austria, Belgium, Denmark, Finland, Germany, Japan, the Netherlands, Norway, Sweden, Switzerland, the United Kingdom, and the United States. These countries feature prominently in the comparative capitalism literature and are categorized as LMEs or CMEs by Hall and Soskice (2001), although some of these countries combine elements of the two ideal types (Kogut & Ragin, 2006). We obtained data for these firms from the Thomson Reuters ASSET4 database, which provides environment, social, and governance data (Hawn & Ioannou, 2016) in the form of over 400 key performance indicators. Limited data availability reduced our sample to 648 firms. Because the relationship between stakeholder engagement and performance differs across industries (Garcia-Castro & Francoeur, 2016; Godfrey, Hatch, & Hansen, 2010), we narrowed our sample of firms to those in the industrials sector according to their ASSET4 classification. We chose this sector because it had the largest number of firms (149, or 23% of all 648 firms). This step enabled us to control for sector effects while retaining institutional diversity. Additionally, the industrials sector is less regulated than other key sectors such as finance and pharmaceuticals and is thus a suitable setting for observing agency in firms' stakeholder engagement. After collecting and matching data to operationalize additional attributes in our configurational model (i.e., firms' internationalization and ownership concentration), our final sample comprised 122 firms for the period from 2004 to 2011, yielding a total of 741 firm-year observations after accounting for missing observations.

| Data and set calibration
The first step in fsQCA is to define sets that represent outcomes (e.g., the set of high-performing firms) and explanatory conditions (e.g., the set of firms with strong policies for protecting shareholder rights). Through a process termed calibration (Ragin, 2000), each case is assigned a degree of membership in each set. We applied the direct method of calibration (Ragin, 2008), which requires the specification of three thresholds based on theoretical and/or empirical knowledge-full membership (1), full nonmembership (0), and a crossover point of maximum membership ambiguity (0.5)-in order to rescale interval variables into fuzzy sets, with the intermediate step of calculating estimates of the log of the odds of full membership. 1

| Outcome: firm performance
Our outcome of interest is firms' financial performance. In accordance with previous studies (Johnson & Greening, 1999;Waddock & Graves, 1997), we operationalized firm performance as return on equity (ROE) and collected ROE data from ASSET4. Although shareholders are ultimately interested in firms' financial returns, it is noteworthy that efforts to attend to shareholders are conceptually and empirically distinct from performance measures such as ROE. As Core, Guay, & Rusticus (2006, p. 658) noted, "shareholder rights can have both negative and positive effects on a firm's operating performance." To calibrate the set of high-performing firms, we used sample-dependent anchors to define set membership, consistent with our explicit interest in these firms' performance relative to others. Specifically, the 80th percentile was the anchor for full membership, the 20th percentile was the anchor for nonmembership, and the median was the crossover point anchor. 2 This is appropriate because for both financial and social dimensions of performance, analysts regularly pinpoint the highest and lowest quintiles to represent high and low performance, respectively (McKnett, 2015;Trunow & Linder, 2015). Because our sample encompasses the period of the 2008 financial crisis, we calibrated the whole sample using different thresholds pre-and postcrisis to adjust for its effects; that is, we calibrated separately for the periods 2004-2007 and 2008-2011. We lagged all 1 Following common practice to avoid theoretical and methodological difficulties of analyzing sets with membership scores of exactly 0.5, we added a 0.001 constant to all such scores (Fiss, 2011;Ragin, 2008). 2 Although our data set comprised the largest sector (industrials) in the ASSET4 data set, we used the entire sample of firms across all sectors in the same 13 countries to inform our calibration thresholds. Because calibration maps absolute values onto a relative scale of membership in a given set, using the larger overall sample for calibration ensures that the sets more accurately depict the underlying performance measures. longitudinal attributes by 1 year compared to performance in order to attenuate reverse causality concerns.

| Stakeholder engagement
We collected data on firms' stakeholder-focused activities from ASSET4. In accordance with our theorizing, we selected attributes reflecting attention to the two central stakeholder groups, employees and shareholders, and operationalized these with two measures for employment quality and shareholder rights from ASSET4, respectively. A firm's emphasis on employment quality is measured by ASSET4 by jointly considering the answers to the questions: "Does the company have a competitive employee benefits policy or ensuring good employee relations within its supply chain?" and "Does the company have a policy for maintaining long term employment growth and stability?" A firm's emphasis on shareholder rights is measured using the question: "Does the company have a policy for ensuring equal treatment of minority shareholders, facilitating shareholder engagement, or limiting the use of anti-takeover devices?" ASSET4 measures weigh and z-score all data points against all companies included in the ASSET4 database. Thus, ASSET4 stakeholder engagement measures are inherently relative, which is appropriate for our analysis because the benefits a firm derives from its stakeholder engagement are relative to, rather than independent from, other firms, who compete for both employees and investment. A firm would score highly if, unlike other firms, it had a strong policy for attending to a focal stakeholder; similarly, a firm's score would be low if, unlike other firms, it did not have a policy for attending to a focal stakeholder. 3 Thus, ASSET4 determines a focal firm's score by comparing its activities to those of all other firms in its industry (for calculating employment quality) or country (for calculating shareholder rights). As above, we calibrated the sets of firms with strong emphasis on shareholder rights and employment quality using sample-anchored thresholds corresponding to the 80th percentile for full membership, median for the crossover point, and 20th percentile for null membership. We again used data for the entire sample and calibrated separately for pre-and postcrisis periods.

| National institutional contexts
We captured institutional context through the strength of a country's capital markets and labor institutions. We captured the strength of labor institutions through two attributes of a country's labor unions-union density and union authority-because unions define norms toward labor by influencing wage negotiations (Checchi & Garcia-Peñalosa, 2010), job security (Atanassov & Kim, 2009), and profitability expectations (Lee & Mas, 2012). We operationalized union density using the ratio of union members to a country's working population (wage and salary earners) in a given year, using data sourced from the OECD.stat database (OECD, n.d.). We operationalized union authority using measures of the degree of authority of unions over workplaces from the Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention, and Social Pacts (Amsterdam Institute for Advanced Labour Studies, 2013). This database includes an index scored from 0 to 1 that captures multiple dimensions of union authority, including control over bargaining, appointment of workplace representatives, strike funds, and enterprise strikes. We directly adopted the union authority index as a set ranging from 0 to 1. Union density was calibrated using sample-anchored thresholds of the 80th percentile, median, and 20th percentile for full membership, the crossover point, and null membership, respectively. We created a higher-order set for a country's labor strength by averaging membership in these two sets.
We operationalized the strength of countries' capital markets by integrating two measures: its stock turnover ratio and its score on the Guillén-Capron minority shareholder protection (GCMSP) index. A country's stock turnover ratio-the value of domestic shares traded divided by their total market capitalization-denotes market strength (Hall & Gingerich, 2009) and defines norms around corporate governance (Jackson & Deeg, 2008). We sourced these data from the World Bank's (n.d.) data catalog. The GCMSP index reflects countries' norms and legal provisions pertaining to minority shareholder rights, focused on the 10 legal provisions considered most relevant to their protection (Guillén & Capron, 2016). Minority shareholder rights are conducive to capital markets and strong capital interests (Aguilera & Jackson, 2003). We obtained GCMSP index data from the cross-national, longitudinal data set of Guillén and Capron (2016). Both measures were calibrated using sample-anchored thresholds of the 80th percentile, median, and 20th percentile for full membership, the crossover point, and null membership, respectively. We then constructed the higher-order set of countries with strong capital markets by averaging the calibrated values for these two sets.

| Need for specialized labor
We captured firms' dependence on specialized labor by classifying firms as high-or low-tech, based on prior findings in the literature that this attribute influences stakeholder engagement (McWilliams & Siegel, 2000;Strike, Gao, & Bansal, 2006). High-tech firms require specialized and high-skilled workers with sufficient talent to invent and commercialize new products, and these firms therefore benefit from initiatives to hire and retain expert employees (Miller et al., 2009). We used Hecker's (1999) classification of industries as high-or low-tech based on Occupational Employment Statistics Survey data on the proportion of employees engaged in scientific, technical, engineering, or R&D roles. We calibrated membership in the set of high-tech firms by matching firms' three-digit Standard Industrial Classification codes to Hecker's classification. This method classified 53 of 122 firms (319 of 741 firm-year observations) as "high-tech." Examples include Boeing, Babcock, and MTU Aero Engines; nonhigh-tech examples in our sample include FedEx, G4S, and Randstad.

| Concentrated ownership
We calculated ownership concentration as the percentage of shares held by the largest shareholder (Thomsen & Pedersen, 2000). We collected ownership data from the OSIRIS database and calculated concentration by considering the largest shareholder's holding (the larger of either direct or total shareholdings), excluding publicly quoted companies and unnamed private shareholders, because these are "considered as unable to exert, as such, control over a company" (Osiris, n.d.). A typical upper threshold for ownership concentration is 20-25%, while a firm with below 5% ownership concentration may be considered management-controlled (Atkinson & Galaskiewicz, 1988;Dam & Scholtens, 2013). Accordingly, we use theoretical anchors corresponding to a single owner's 25% shareholding for full membership, 15% for the crossover point, and 5% for null membership.

| Internationalization
Following prior studies (Sullivan, 1994), we measured a firm's internationalization as its foreign sales as a percentage of total sales. We collected this data from the Worldscope database. We calibrated this attribute using sample dependent anchors corresponding to the 80th percentile for full membership, median for the crossover point, and 20th percentile for null membership. Table 2 presents descriptive statistics and correlations for all variables used to calibrate attribute sets. Table 3 reports descriptive statistics of the cross-country case distribution and countries' mean institutional scores in our observation period. The distribution of countries on scales of capital and labor, created by averaging standardized values of GCMSP and stock turnover ratio for capital and union density and union authority for labor, is depicted in Figure 2. Of the 741 total observations, 226 represent firms in countries typically classified as LMEs (Australia, United States, and United Kingdom), and 515 represent firms in countries typically classified as CMEs (Austria, Belgium, Denmark, Finland, Germany, Japan, the Netherlands, Norway, Sweden, and Switzerland). However, as Table 3 and Figure 2 show, our operationalization goes beyond a binary dichotomy of LME/CME. For instance, Finland, Germany, and Sweden display features of a hybrid context with strong labor institutions and comparatively strong capital institutions. Contexts with strong capital and weak labor institutions include not only the United States and United Kingdom but also, unexpectedly, Japan.

| Analysis
Prior to conducting sufficiency analysis, we conducted necessity analyses to explore whether any condition in our model was by itself necessary in order to obtain high financial performance. A condition is necessary if it must be present for the outcome to occur (Ragin, 2008). Consistency thresholds for necessity analyses are typically pegged above 0.90 (Greckhamer, Furnari, Fiss, & Aguilera, 2018;Schneider & Wagemann, 2012), and a condition may be deemed necessary if it has both high consistency and high coverage scores.
We conducted our sufficiency analysis using Ragin's (2008) fsQCA truth-table approach, which constructs and analyzes a data matrix referred to as a "truth table" that includes 2 k rows (k = number of explanatory conditions; i.e., 2 7 = 128). Each row represents a logically possible combination of conditions. Using Boolean algebra, the truth table is reduced into a simplified expression of combinations linked to the outcome (Ragin, 2008). For this purpose, we set two criteria: (a) a frequency threshold for the minimum number of cases that must belong to a combination for it to be considered by the analysis and (b) a consistency benchmark to identify combinations that are reliably linked to an outcome. We followed recommended levels (Ragin, 2008) in choosing a raw consistency threshold of 0.85, supplemented by a 0.65 proportional reduction in consistency (PRI) threshold following prior research (Greckhamer, 2016). We chose a frequency cut-off of two cases, recommended by Ragin (2008) for relatively smaller samples, but also run the analyses with alternative thresholds. As we document in Appendix 1, varying frequency and consistency thresholds to alternative levels within the recommended bounds produced similar results to those presented here. We used the fsQCA 3.0 software (Ragin & Davey, 2016) to conduct our analyses.
It is recommended that, upon completion of QCA analyses, researchers should return to cases to interpret the essence of identified configurations (Greckhamer et al., 2018). Accordingly, we collected qualitative data for cases representing each of the five configurations identified in our main analysis. We performed a content analysis of these cases using all available Factiva sources and firm annual reports for the observation period that mentioned the firm and contained the word(s) "employee(s)," "shareholder(s)," and/or "stakeholder(s)." This yielded  9,910 articles. In the first stage, our analysis was by firm. In the second stage, two researchers assessed commonalities between those firms representing the same configuration.

| RESULTS
Necessity analyses found that neither the presence nor absence of any of the seven conditions in our model is necessary for high performance, which reinforces the expectation of complex causality. Table 4 presents the results of sufficiency analysis for high financial performance. We follow the convention of using a configuration chart to report a combination of parsimonious and intermediate solutions 4 (Ragin & Fiss, 2008 Large circles indicate core conditions, whereas small circles indicate peripheral conditions. To aid the interpretation of findings reported in Table 4, we summarize our qualitative insights in Table 5 and incorporate them in our discussion of the results. Table 4 shows that five configurations of stakeholder engagement, national institutions, and firm attributes are consistently linked to high financial performance. These include substitutionary (Configurations 1 and 2), minimalist (Configurations 3 and 5), and encompassing (Configuration 4) stakeholder engagement strategies. No configuration links complementary engagement strategy to high performance. Specifically, Configurations 1 and 2 show that a substitutionary engagement strategy is consistently associated with high performance in contexts characterized by strong capital institutions (and, for Configuration 2 only, weak labor institutions). Configuration 3 represents a minimalist engagement strategy in a context of strong capital and labor institutions, whereas Configuration 4 represents an encompassing engagement strategy in the same context but pursued by different kinds of firms. Finally, Configuration 5 represents a minimalist engagement strategy in a context with strong capital institutions only.
Configuration 1, which we label as multinational, employee-first firms, is consistent with a substitutionary engagement strategy. Engagement focused on employees rather than shareholders is linked to high performance for highly internationalized firms headquartered in countries with strong capital institutions. This configuration is consistent with firms transcending normative expectations in their home countries in order to incentivize skilled workers to join and stay with them. In contrast, the institutional context in their home country already protects investors, and firms in this configuration choose not to undertake further shareholder engagement activities. As indicated in Table 5, some exemplar firms representing this configurationsuch as the successful conglomerates 3M (based in the U.S.) and Smiths Group (based in the 4 The QCA's truth table algorithm for sufficiency analyses produces distinct parsimonious and intermediate solutions. Intermediate solutions integrate simplifying assumptions that are consistent with empirical evidence at hand and with existing knowledge regarding single conditions that compose logical remainders (i.e., "easy" counterfactuals); parsimonious solutions may integrate both easy counterfactuals as well as difficult ones (i.e., those that are consistent with the empirical evidence but not with theoretical knowledge; Fiss, 2011). 5 The core-peripheral distinction builds on prior work in strategy and organization studies (e.g., Grandori & Furnari, 2008;Hannan & Freeman, 1984) that has conceptualized core features as central features in an organization that are surrounded by peripheral features that reinforce the core. Accordingly, QCA scholars view core conditions as those with the strongest evidence linking them to the outcome of interest while peripheral conditions are those that are linked to the outcome through weaker evidence (Fiss, 2011). U.K.)-have a history of conflict with activist investors, whose attempts to drive strategy are resisted by these firms' executives.
Similar to Configuration 1, Configuration 2, which we label as balancing interests, underscores the tensions among stakeholders that link substitutionary engagement to high performance. In a context of strong capital but weak labor institutions, for high-tech firms, focusing on employees rather than shareholders is linked to high performance. Where formal institutions do not support labor, our findings are consistent with the idea that high employment quality can facilitate retaining and motivating employees, especially for high-tech firms dependent on specialized labor. Conversely, as shareholder rights are already protected in this context, further shareholder-focused activities may not provide differential advantage. Qualitative evidence shows that some exemplar firms representing this configuration-for example, the U.S.-based Boeing and Raytheon, and the U.K.-based Babcock-perceive extensive pressure from employees and struggle to balance attention to employees on the one hand and other stakeholders such as shareholders and large government clients on the other (Table 5).
Configurations 3 and 4 both capture firms in hybrid institutional contexts that combine strong labor and capital institutions, which implies high expectations from different stakeholders toward firms (Aguilera & Jackson, 2003). However, they represent two distinct stakeholder engagement strategies. Configuration 3, which we label as multinational Nordic players, represents minimalist engagement, because it does not include the presence of any stakeholder engagement activities. The combination of strong capital and labor institutions is found in some European countries, such as Finland, Sweden, and Germany. In these hybrid contexts, "employee turnover ran in excess of 20% per year. The first step in their transformation was coming up with a description of desired behaviors," with 42% profit growth in 4 years (HR Magazine, July 1, 2011).
• Executives were confident that content employees ultimately produced satisfied shareholders (e.g., Accenture, as covered in Financial Express, December 7, 2009).
• "People are our greatest asset, and attracting, nurturing and developing the best talent remains a top priority," said LaMae Allen deJongh, Managing Director, US Human Capital and Diversity, Accenture.
"We are honored to be recognized once again as a leading employer of choice-particularly as we continue to actively recruit for a variety of skills in the United States and around the world" (Business Wire, 2010).

Reluctance to let investor demands drive strategy
• The firms in this configuration have often adopted a stance against activist investors, believing that their proposals are not conducive to long-term growth.
• "Smiths Group has been unloved in the City for years, frequently criticized as an outdated mini-conglomerate with disparate businesses that don't hang together" (Observer, March 22, 2009).
• Activist investors have raised agency concerns (e.g., "Smiths Group shareholders rebel over pay," Financial Times, November 23, 2011). • The government, a key stakeholder for some firms in this configuration, often represented the interests of employees vis-à-vis management. For example, by pushing for higher salaries (U.S. Federal News Service, May 12, 2011). In the United Kingdom, for instance, "remuneration at companies such as Capita is sensitive given their reliance on contracts with central and local government, which are enduring budget cuts" (Financial Times, April 5, 2011). Difficulty of simultaneously pleasing multiple stakeholders, especially shareholders and government customers

Configuration
• "Pay by performance is notoriously tricky to get right in boardrooms-but in prisons? That was the task when it came to drawing up Serco's contract at HMP Doncaster. In the first deal of its kind in the UK, 10% of payment for the first 4 years of the pilot depends on the prison operator cutting reoffending rates. It looks less than ideal […] If the contract were structured differently, so that results really affected payments, then it could reach a point at which a group-and perhaps even its shareholders-became reluctant to take on this work, because so much money would depend on factors beyond their control" (Financial Times, November 25, 2011).
• to the financial crisis) activist equity fund Cevian saying on Wednesday it controlled 5% of the voting shares in the truck maker and would pressure the firm to disburse some of its cash pile to shareholders" (Financial Times, September 62,006).
• "Shareholder pressure on Volvo, the world's second largest truck maker, intensified yesterday when a Swedish fund management company added its voice to calls to restructure its balance sheet. Mats Qviberg, chief executive of Oresund, an asset manager with a 1% stake in Volvo, said Volvo should offload assets and streamline its balance sheet" (Financial Times, September 7, 2006).
• Gea Group was similarly involved in legal disputes with shareholders (German News Digest, November 2, 2009). Reduced incentive to focus on minority shareholders as investor pool widens • MTU AeroEngines had less incentive to focus on minority shareholders as it tried to attract a shareholder to take a large stake in the company: "MTU is on the lookout for a core shareholder in order to avert a potential hostile takeover from abroad. The core investor should take a double-digit percentage stake in MTU's shares which are currently floating freely, Behle said" (German News Digest, November 27, 2009).
• "A growing number of financial investors is showing an interest in SGL Carbon, the German graphite specialist" (Frankfurter Allgemeine Zeitung, August 14, 2007).

Integrating stakeholder interests
Firms that highlight positive synergies between focusing on employees and shareholders, and implement initiatives to align their interests Encompassing Konecranes (after the financial crisis) Successful initiatives to align interests of employees and shareholders • Konecranes implemented several measures focusing on employees and shareholders, aiming to align their interests and highlight synergies between their needs.
• A key measure included providing stock options to employees: "The Company has issued stock option plans for its key employees, including top and middle management, and employees in certain expert positions. The purpose of the option schemes is to motivate key personnel to contribute to the long-term success of the Company and to create a common understanding of and commitment to the creation of shareholder value" (Konecranes Annual Report, 2009).

T A B L E 5 (Continued)
Configuration Engagement strategy Exemplar firms Qualitative evidence • A high-tech company, Konecranes articulated employee acquisition and retention as a significant risk in a competitive environment and took several steps to focus on employees: "Konecranes' ability to operate is dependent on the availability, capabilities, and expertise of professional personnel. The ability to recruit and retain personnel is of key importance for the future success of the company […] The continuing retirement of skilled personnel challenged Konecranes to enhance the transfer of tacit knowledge and skills to younger professionals. Konecranes continued to invest in training and enhanced the availability and communication of its training options" (Konecranes Annual Report, 2009). employees' and shareholders' interests are already reasonably well protected by institutions, which potentially restricts the benefits of implementing further policies to serve these stakeholders. Here, not focusing on activities aimed at shareholder rights is connected to high performance for high-tech firms (a core condition) that are highly internationalized (a peripheral condition). Additional qualitative evidence on the cases exemplifying this configuration shows that some of these firms issue dual-class stock and rely on cross-shareholdings with strategic partners (Table 5). Also, some of these firms have important government (e.g., Metso) or family (e.g., Kone) ownership, which may insulate them from pressure to cater to short-term investors or minority shareholders. In contrast, Configuration 4, which we label as integrating stakeholder interests, represents an encompassing stakeholder engagement strategy. In hybrid institutional contexts, for hightech firms (a peripheral condition) that lack concentrated ownership and internationalization, a strategy of integrating shareholders' (peripheral condition) and employees' (core condition) interests is linked to high performance. Qualitative evidence from Konecranes reinforces this idea. This high-tech firm adopted multiple measures to align the interests of shareholders and employees, such as providing stock options to employees, shifting to a two-tier management structure, and investing in employee-focused training and retention activities. Konecranes articulated the importance of focusing on employees to boost its competitive advantage and shortterm performance, which is also important for attracting minority shareholders (Table 5). Further comparing Configurations 3 and 4 suggests that in the hybrid institutional contexts they represent, for international high-tech firms (i.e., Configuration 3), a minimalist strategy is consistently linked to high performance; this may be the case because these firms may be shielded from public scrutiny at home. On the other hand, an encompassing strategy that actively attends to the rights of a broad set of stakeholders is linked to performance for domestically focused high-tech firms with dispersed ownership that are exposed to local stakeholders' scrutiny (i.e., Configuration 4).
Configuration 5, which we label as partnering to overcome stakeholder pressures, represents a minimalist engagement strategy. In a context of strong capital institutions, not investing resources to engage with either employees or shareholders is consistently linked to high performance for high-tech firms with concentrated ownership and low internationalization. This configuration represents firms headquartered in Sweden and in Japan. Notably, the exemplar firms belonging to this configuration often engage in strategic partnerships, including equity alliances. For example, Volkswagen has a large equity stake in Scania (Sweden), and Isuzu (Japan) collaborates closely with Toyota. Together, these features potentially limit the exemplar firms' exposure to pressures from labor and shareholders because they give the firms wider access to skills and capital. Table 4 represents stakeholder engagement strategies that are consistently linked to high financial performance across institutional contexts. We complemented these findings through subsample analyses to identify the configurations linked to high financial performance within LME, CME, and hybrid economies. Subsample analyses are useful in configurational studies for developing insights about performance within rather than across contexts (Greckhamer, Misangyi, Elms, & Lacey, 2008). For this purpose, we used data for country-institutions (see Table 3 and Figure 2) to split our sample into four subsamples as follows: (a) countries with strong capital but weak labor institutions; (b) countries with strong labor but weak capital institutions; (c) countries with strong labor and capital institutions; and, finally, and (d) countries with weak labor and capital institutions.

| Subsample analyses
To analyze these subsamples, we included all attributes except those representing the strength of labor and capital institutions, which were used to create the subsamples. For each analysis, we chose a raw consistency threshold of 0.80, a PRI consistency threshold of 0.65, and a frequency threshold of two cases, which corresponded to the inclusion of at least 96% of cases. We report results for these analyses in Tables 6-9.  Table 6 shows the results for firms in countries with strong capital and weak labor institutions (including the United States, United Kingdom, Australia, and Japan). We identify five configurations consistently linked to high performance. As in the main results presented in Table 4, both minimalist (Configurations 1 and 3) and substitutionary (Configuration 2) stakeholder strategies are associated with high performance in this context in combination with some firm-level attributes. In addition, an encompassing stakeholder engagement strategy focused on both shareholders and employees is linked to high performance for internationalized low-tech firms with concentrated ownership (Configuration 4) or domestically-oriented high-tech firms (Configuration 5). Table 7 shows the results for firms in countries with strong labor and weak capital institutions (including Austria, Belgium, Denmark, the Netherlands, and Norway). This solution contains only one configuration and a low overall coverage, which may suggest that the diversity of institutional environments within the subsample and/or the diversity of firms in this subsample is not represented as well by the included dimensions as are the other subsamples. In other words, there may be various paths to high performance in these countries, but only one is consistently linked to high performance. The single configuration in these results indicates that a T A B L E 6 Configurations for achieving high performance for sub-sample with strong capital and weak labor institutions (including United States, United Kingdom, Australia, and Japan) lack of shareholder engagement is connected to high performance for low-tech firms that lack ownership concentration and are not internationalized. Table 8 shows the results for firms in countries with strong labor and capital institutions (including Sweden, Finland, and Germany). This subsample analysis identified four configurations that are consistently linked to high performance in this institutional context. As in the main analysis, two configurations correspond to the minimalist stakeholder strategy (Configurations 1 and 2), and one corresponds to an encompassing stakeholder strategy (Configuration 4). In addition, Configuration 3 indicates that attention to shareholders but not to employees (peripheral conditions) is linked to performance for some firms with dispersed ownership. This configuration represents a partially complementary stakeholder engagement approach that complements the strong capital institutions (which have recently become strong) while not complementing the traditionally strong labor institutions.
Finally, Table 9 shows the results for firms in an institutional environment that combines both weak capital and labor institutions (i.e., Switzerland). We found four configurations consistently linked to high performance in this subsample. However, this subsample includes only one country, hence we interpret these results with caution as to their generalizability to other countries with a similar institutional environment. These findings suggest that a stakeholder engagement strategy focusing on employees but not on shareholders is linked to performance in this context (Configuration 3), in combination with high internationalization and a low dependence on specialized labor (both peripheral conditions). For firms that have a concentrated ownership base and a high dependence on specialized labor, focusing on shareholder rights and employment quality (both peripheral conditions) is linked to high performance (Configuration 4).
T A B L E 7 Configurations for achieving high performance for sub-sample with strong labor and weak capital institutions (including Austria, Belgium, Denmark, the Netherlands, and Norway) T A B L E 8 Configurations for achieving high performance for sub-sample with both strong capital and labor institutions (including Sweden, Finland, and Germany) T A B L E 9 Configurations for achieving high performance for sub-sample with both weak capital and labor institutions (including Switzerland)

| Additional analyses
We performed sensitivity analyses for different calibration thresholds to those used in our main analysis. We altered the anchors for full and null membership in a set from the 80th and 20th percentiles to the 90th and 10th percentiles. As reported in Appendix 2, the results show minor changes related to the number of configurations and solution coverage. We continue to observe configurations that correspond to the substitutionary and minimalist engagement strategies. As in the main analysis, we do not observe a complementary strategy linked to high performance. We also do not observe an encompassing strategy linked to high performance; in our main analysis, this strategy recorded the lowest coverage. Finally, fsQCA does not assume causal symmetry. Accordingly, in Appendix 3 we report the results of sufficiency analysis for the absence of high financial performance and discuss the three configurations we found to be consistently linked to this outcome; we also present an exemplar firm representing each of these configurations.

| DISCUSSION
The primary contribution of this study is to advance a configurational perspective on stakeholder engagement. Prior research documents the effects on performance of providing better governance and better treatment of employees, respectively. However, by focusing on either of these effects, often only in any one context, prior studies are unable to account for differences in the relative power of labor and capital across nations. In this study, we develop a typology of stakeholder engagement strategies to map different approaches firms can use to generate financial performance through engagement with labor and shareholders in different institutional environments. Our typology proposes that firms can pursue stakeholder engagement in a manner that is complementary, substitutionary, minimalist, or encompassing given their respective institutional contexts (see Table 1). Guided by this typology, our analysis helps to unravel the complexity underlying the link between performance and stakeholder engagement.
Crucially, we found that different engagement strategies are linked to high performance in different contexts. In contexts with strong capital and weak labor institutions, a substitutionary strategy is most salient in configurations consistently linked to high performance. In contrast, in hybrid contexts characterized by both strong labor and capital institutions, under different conditions, the extremes of minimalist and encompassing strategies are consistently linked to high performance. Moreover, in our main analysis, we found that no configuration that is consistently linked to high performance entails a complementary stakeholder engagement strategy (see Table 4), which would focus on attending to the particular stakeholder group represented by strong local institutions. This could be because institutionally supported stakeholders already are in strong positions to get their demands fulfilled, a strategy based on focusing on those stakeholders provides the least opportunity for successful differentiation, which reinforces the idea that for stakeholder engagement more is not always better (Harrison & Bosse, 2013). Our subsample analyses further identify the contexts in which specific strategies are consistently linked to high performance.
From a strategic management perspective that emphasizes firms' desire to translate stakeholder engagement into economic value (Wang, Choi, & Li, 2008), our findings imply that any strategy may provide different value across different contexts because its results are dependent on a combination of institutional and firm-level factors. Thus, the theoretical and empirical connections we forge between national institutions, stakeholder engagement, and financial performance advance research at the interface of strategy and stakeholder theory (Tantalo & Priem, 2016). Our configurational model informs efforts to disentangle the conditions under which different forms of stakeholder engagement are linked to financial returns for firms (Barnett, 2007;McWilliams & Siegel, 2001) across different contexts. Additionally, our qualitative results (Table 5) provide nuance as to how different factors within a firm may combine to create opportunities to benefit from stakeholder engagement. Similarities across firms in a particular configuration, captured by labels derived from the qualitative analysis, present new pathways for scholars to explore at the intersection of strategy and stakeholder theory.
More broadly, firms' strategies to navigate tensions between labor and capital have implications for firms' efforts to balance the need for conformity-such as when they give attention to a locally dominant stakeholder group-with the need to be distinct, and, consequently, for their performance. Each engagement strategy in our typology represents a different approach to managing tensions between conformity and differentiation. Research on comparative institutions highlights the benefits accruing to coherent firm-level strategies appropriate to local contexts (Hall & Soskice, 2001). From this perspective, there are benefits to conformity. However, although coherence between the elements of an institutional system is consistent with comparative advantage (Witt & Jackson, 2016), competitive advantage requires firms to differentiate themselves from competitors. From this perspective, there are benefits to being distinct. Prior research suggests striking an optimal balance between conformity and distinctiveness (Zhao et al., 2017). Our typology and empirical investigation imply that this balance varies across firms and contexts, and that, in many instances, firms benefit from emphasizing distinctiveness rather than conformity, as encapsulated by the substitutionary engagement strategy.
Finally, our configurational perspective stands in contrast to much of the cross-national literature on governance insofar as it takes account of institutional and firm-level factors, as well as performance differences across firms. The varieties of capitalism literature emphasizes national political economies as the central determinants of how firms address coordination problems. Though the ambition of the varieties of capitalism approach at its outset was to be "actor-centered" in examining how firms address coordination problems (Hall & Soskice, 2001, p. 6), the actor-centered perspective has remained underdeveloped. By acknowledging the diverse strategies that exist within the same context, we underscore the relevance of actorcentered explanations within the varieties of capitalism approach.

| Directions for future research
One direction for future research is suggested by our finding that substitutionary stakeholder engagement is more salient than complementary engagement in configurations that are consistently linked to high performance (as indicated by the high coverage scores of Configurations 1 and 2 and the absence of any configuration representing a complementary strategy in our solution). To date, the notion that firms substitute for weak institutions has been primarily applied to institutional voids in emerging markets (Marquis & Raynard, 2015) that denote transaction hazards due to opportunism and ineffective protection of property rights (Khanna & Palepu, 2000). By extending the logic of institutional voids to the developed countries in our study, we highlight that institutional systems develop in a way that privileges certain key actors (e.g., shareholders and equity-based financing in LMEs; banks and labor in CMEs), and thus develop around locally-accepted answers to the question: "Who or what counts?" These expectations may create voids that leave other stakeholders' demands unmet and pose strategic challenges and opportunities for firms.
As briefly noted above, a configurational model is limited in the number of attributes it can include, due to the viable balance between model conditions and sample size (Marx & Dusa, 2011) and the exponentially increasing complexity of a study's model and thus potentially of its findings (Greckhamer et al., 2018). Future research could build on our study by exploring further contingencies that explain the stakeholder engagement-financial performance relationship. Such research may consider that some firms with exceptionally high or low status have greater leeway to pursue a substitutionary engagement strategy (Phillips & Zuckerman, 2001). To give another example, future research could take a different approach to ownership-related attributes. Building on prior research (e.g., Thomsen & Pedersen, 2000), such research could go beyond our consideration of ownership concentration by considering the importance of ownership identities; indeed, qualitative evidence for Configuration 3 presented in this study suggests that ownership identity may be relevant to the link between stakeholder engagement and high performance for some firms.
Moreover, considering that returns from stakeholder engagement differ by industry (Flammer, 2015), future research should further investigate industry-level contingencies that shape firms' ability to capture value from their stakeholder engagement strategies. For example, the natural resources sector is subject to substantial scrutiny, and its firms have particular need for support from social actors (see Henisz, Dorobantu, & Nartey, 2014). Future research may also consider other important stakeholders, such as firms' customers, stakeholders representing the natural environment, and the state (Kollman & Prakash, 2001;Ortiz-de-Mandojana, Aguilera-Caracuel, & Morales-Raya, 2014). Because the state shapes labor and capital institutions, it is important to examine whether firms' stakeholder engagement (e.g., attending to employee concerns in a context of strong capital institutions) engenders a reaction from the state.
National institutional systems are themselves constellations of factors that may be subject to change and shaped by firms over time (Henisz & Zelner, 2003;Hillman, Keim, & Schuler, 2004). As firms in CMEs begin to give greater privilege to shareholder rights, norms relating to the appropriate standing of shareholders might change. Our analysis considers annual changes in institutional contexts, but only over a relatively limited time period. We see potential value in exploiting exogenous shocks as quasi-experiments in order to advance research in this domain. For example, an increase in takeover protection has allowed U.S. managers to channel greater attention to nonshareholding stakeholders (Kacperczyk, 2009). Accordingly, the increase in discretion stemming from lower danger of being disciplined by markets may foster a substitutionary engagement strategy. Future research could assess what kinds of firms use this newly gained agency to adjust their stakeholder engagement activities. The advantages of this approach notwithstanding, appropriate exogenous shocks that apply across different countries are likely to be relatively rare, hence, this approach primarily lends itself to within-country studies.
In this study we have adopted a configurational theoretical approach because it has the capacity to address patterns across a variety of contexts and practices. FsQCA is well suited to study complex phenomena such as how institution-and firm-level conditions in combination are linked to superior performance. It is also suitable for identifying equifinal relationships, such as when multiple stakeholder engagement strategies are associated with similarly superior performance. Taking complexity and equifinality seriously, we used fsQCA to explore how institutional contexts and firm-level factors combine with firms' stakeholder engagement activities to create patterns linked to high performance outcomes.
However, we also note that even though a configurational approach such as QCA offers benefits in helping scholars understand complex questions, they are not without limitations. In particular, researchers should recognize that findings form QCA analyses provide empirical evidence of association and accordingly they should exercise caution and theoretical rigor in drawing causal inferences based on these findings from QCA analyses. Moreover, while reverse causality and omitted variable bias are issues common to several analytical techniques, we note that conventional quantitative methods have more formal tools to deal with them than QCA. To build stronger evidence of causal relationships, future research may also complement our cross-context configurational approach by performing in-depth qualitative within-case analysis (Schneider & Rohlfing, 2013, 2016.