Introduction

The entry of a new business or type of business into a society raises a fundamental ethical question: Will this be good for our society (Beauchamp and Bowie 2001; Wicks and Freeman 1998)? Individuals implicitly or explicitly answer this question as part of their many roles as citizens, members of families and communities, consumers, and employees. A type of business entering new countries in the last decade is multinational enterprises from emerging markets (EMNEs), such as the entry of Chint Solar, a solar energy company from China, into The Netherlands. Will this company provide good things to Dutch society, such as products that reduce greenhouse gas emissions and employment practices that are consistent with Dutch labor standards?

The concept of organizational legitimacy can be used to answer these questions. Legitimacy represents the appropriateness of organizational purposes and actions to stakeholders in a particular society (Deephouse et al. 2017; Suchman 1995; Tost 2011; Zimmerman and Zeitz 2002). Research in both organization studies and business ethics indicates that gaining legitimacy enables organizations to survive and thrive (Palazzo and Scherer 2006; Suchman 1995). Thus, international business research recognizes the importance of gaining legitimacy for EMNEs entering a new host country because the EMNEs suffer from the liabilities of foreignness and emergingness in the host country (Buckley and Casson 1976; Madhok and Keyhani 2012; Zaheer 1995).

However, there is little research examining how host countries evaluate the legitimacy of EMNEs, especially individuals in host countries who are increasingly recognized as the micro-foundations of legitimacy (Bitektine and Haack 2015; Deephouse et al. 2017; Suddaby et al. 2017; Tost 2011). Extant research offers different theorizations and measures of legitimacy with only limited applicability to EMNEs, an organizational form prominent in international business research; for example, Panwar et al. (2014) considered only social and environmental criteria of forest products companies, and Bitektine et al. (2020) only considered sociopolitical and cognitive dimensions. Measuring individuals’ perceptions of the legitimacy of EMNEs in host countries would enable researchers, policy-makers, and activists to evaluate the legitimacy of EMNEs in their societies and managers of EMNEs to assess the efficacy of legitimacy-enhancing strategies, such as corporate social responsibility (CSR) and institutional isomorphism (Du and Vieira 2012; Kostova and Zaheer 1999; Miller and Eden 2006; Wu and Salomon 2016). Such work would also contribute to the “ongoing process” of developing validated instruments in psychology and management (Flake et al. 2017, p. 374; Schwab 1980), including legitimacy. Haack and Sieweke (2020, p. 153) in their commentary accompanying Bitektine et al. (2020), express their “understanding that progress in social evaluation research (which includes legitimacy) requires the concerted effort of many researchers over many years (Deephouse et al. 2017; Deephouse and Suchman 2008).”

A second gap is the limited attention to how legitimacy is related to theories of business ethics; a notable exception is Palazzo and Scherer (2006). Organizations are systems of people, processes, and structures with goals and means that acquire inputs from suppliers and provide outputs of goods or services to customers, all of whom are embedded in a society having moral values, legal rules, and common understandings (Scott 1995). Organizational legitimacy is the appropriateness of organizational purposes, processes, and outputs to stakeholders in society (Brønn and Vidaver-Cohen 2009; Suchman 1995; Tost 2011). This attention to evaluating both ends and means is within the domain of business ethics (Wicks and Freeman 1998, p. 124).

We address these two gaps in the following way. We first put an ethical foundation under the conceptualization of legitimacy, focusing on deontology, utilitarianism, and the new pragmatism. We then draw from information processing and legitimation theories to elaborate how individuals make legitimacy judgments (e.g., Bitektine 2011; Ehrhart and Ziegert 2005; Tost 2011). Using these foundations, we develop and validate a measure of individuals’ perceptions of the legitimacy of EMNEs.

Our empirical setting focuses on Dutch individuals evaluating the legitimacy of Chinese EMNEs in The Netherlands. In the context of the new pragmatism in organization studies, the problematic question people seek to answer is: Are Chinese EMNEs entering the Netherlands good for our society, communities, and individual people in our communities—that is, do Chinese EMNEs “count(s) as good or bad organizing?” (Wicks and Freeman 1998, p. 132). We develop a measurement scale for legitimacy in this context, and analysis of survey data using PLS-SEM provides evidence of construct validity.

Our study makes several contributions. First, putting an ethical foundation under the conceptualization of legitimacy shows that legitimacy is a fundamentally an ethical concept invites further application of ethical theorizing to legitimacy with cascading influence into the variety of literatures where legitimacy is applied, such as entrepreneurship (Aldrich and Fiol 1994), institutional theory (Deephouse et al. 2017), and international business (Kostova and Zaheer 1999).

Many theorists proposed detailed models of the legitimacy judgment process (e.g., Bitektine 2011; Bitektine and Haack 2015; Tost 2011), but there is a lack of empirical research opening up the “black box”; a notable exception is Jacqueminet and Durand (2020, p. 9). Our next contribution is formulating and empirically testing a model of the legitimacy judgment process derived from past research. In so doing, we define validating knowledge as an individual-level construct indicating the resulting information retained after filtering the validity cues provided by field-level institutions for use in propriety judgments (Bitektine and Haack 2015; Cyert and March 1963; March and Simon 1958; Tost 2011). We find support for this process linking these components: validating knowledge → propriety judgments → generalized judgment, thus improving past theorizing (e.g., Bitektine 2011; Ehrhart and Ziegert 2005; Tost 2011).

Our research also informs research on the dimensionalization of individual perceptions of legitimacy. Rather than specifying ex ante a particular set of dimensions to measure (e.g., Alexiou and Wiggins 2019; Bitektine et al. 2020), we cast a wide net and include indicators of all types in our initial item list before reducing and purifying the list. Our final process model specifies six dimensions. Consistent with Suchman (1995), Tost (2011), and Jacqueminet and Durand (2020), a two-item measure of generalized legitimacy emerged as a separate factor. Propriety judgments consisted of pragmatic legitimacy and sociopolitical legitimacy. The latter emerged when measures of moral and regulatory legitimacy loaded on the same factor; this is consistent with deontological ethics (Beauchamp and Bowie 2001; Chakrabarty and Bass 2015), positive norm connectedness (Bascle 2016), entrepreneurship research (Aldrich and Fiol 1994), and the recent study by Bitektine et al. (2020) but contrary to others who view moral and regulatory dimensions as distinct (Deephouse et al. 2017; Scott 2014; Suchman 1995). The emergence of pragmatic legitimacy, initially theorized by Suchman (1995), is consistent Foreman and Whetten (2002) at the form level and with the recent scale developed by Alexiou and Wiggins (2019) at the organizational level.

Validating knowledge has three dimensions: level of knowledge; challenging knowledge; and endorsing knowledge. In contrast to research at the collective level that combined endorsing and challenging media reports in one measure (Bansal and Clelland 2004; Deephouse 1996), separating them at the individual level improved the fit of the model dramatically and is consistent with research claiming that good and bad news have asymmetric impacts (Baumeister et al. 2001; Haack et al. 2014).

Taken together, future research could adapt our validated measure to other home country–host country settings to assess the ethical behavior of EMNEs. More generally, this research contributes to the “ongoing process” of developing better measures of individual perceptions of legitimacy (Flake et al. 2017, p. 374; Haack and Sieweke 2020).

Our paper is structured as follows. We first review research measuring individual perceptions of legitimacy. Our theory section integrates research on individual perceptions of legitimacy with business ethics and international business research, yielding an ethically justified conceptualization of individual perceptions of the legitimacy of EMNEs in a host country. Our next section develops survey items to measure legitimacy based on this conceptualization, items used in past research, and our particular context, Dutch citizens evaluating Chinese EMNEs. Using the refined scale and validated constructs, we then examine the relationships among the dimensions and two individual-level outcomes. Finally, we discuss our implications and limitations.

Empirical Research on Individual’s Perceptions of Legitimacy: A Growth Industry

Research on individuals’ perceptions of legitimacy started slowly, with Elsbach’s (1994) examination of the normative legitimacy of the California cattle industry followed Foreman and Whetten’s (2002) examination of the pragmatic and cognitive legitimacy of Illinois cooperatives. Since 2012, this research has accelerated. There are many attempts to create measures that are both faithful to contexts and generalizable across contexts, including two very recent nearly simultaneous construct validation studies with different theorizations, samples, and results (Alexiou and Wiggins 2019; Bitektine et al. 2020).

We offer a more comprehensive review that includes not only the five articles on industries or organizations identified in the reviews of Alexiou and Wiggins (2019) and Bitektine et al. (2020) but also 10 others. Our first consideration is the legitimacy subjects in these studies: organizations (Bachmann and Ingenhoff 2016; Bitektine et al. 2020; Choi and Shepherd 2005; Díez-Martín et al. 2013; Elsbach 1994; Jahn et al. 2020; Kim et al. 2014; Pollack et al. 2012; Shepherd and Zacharakis 2003); industries, forms, or sectors (Finch et al. 2015; Foreman and Whetten 2002; Gau 2013; Humphreys and Latour 2013); practices (Jacqueminet and Durand 2020; Panwar et al. 2014; Yang et al. 2012); activities (Streib and Poister 1999); and brands (Guo et al. 2017).

Second, legitimacy is theorized in a variety of ways. Many studies in business ethics focus on the legitimacy of specific practices or types of organizations on specific dimensions, such as the moral legitimacy of CSR practices (Panwar et al. 2014) or credibility and communication in the energy sector (Finch et al. 2015; Jahn et al. 2020). The construct validation effort by Bitektine et al. (2020) selects two dimensions for scale development, cognitive legitimacy, and sociopolitical legitimacy. This dimensionalization was promulgated by Aldrich and Fiol (1994) and adopted in Bitektine (2011). After their review of past literature, Alexiou and Wiggins (2019 p. 472) select the prominent and well-cited dimensionalization of Suchman (1995): pragmatic, moral, and cognitive; “while these dimensions by no means exhaust all possible types of legitimacy, they capture the majority of the conceptual domain of the construct and represent the most frequently measured dimensions of legitimacy (Díez-Martín et al. 2013; Elsbach 1994; Foreman and Whetten 2002; Humphreys and Latour 2013; Pollack et al. 2012).” Moreover, few papers examined a multi-step process of legitimacy judgment (e.g., Finch et al. 2015). Thus, the number and naming of specific dimensions remains an unresolved issue. The selection of dimensions influences the nature of the measurement instruments.

Third, we consider the theorized antecedents and outcomes of legitimacy, and there is a wide variety in past research. For instance, Panwar et al. (2014) tested the influence of family versus public ownership on the legitimacy of CSR actions in the forestry industry, and Jacqueminet and Durand (2020) examined the influence of validity and propriety dimensions on the implementation of CSR practices by subsidiary managers of a large multinational energy company. Alexiou and Wiggins (2019) and Bitektine et al. (2020) undertake explicit criterion-related validation. The former use an imaginary scenario experiment in which Amazon MTurk respondents could contribute up to $100 in a crowdfunding campaign for a microfinance company. Bitektine et al. (2020) use brand loyalty of students in one university to three different taxi companies competing in Montreal. These two outcomes have low commitment and low switching costs. In contrast, implementation of CSR practices by subsidiary managers in Jacqueminet and Durand (2020) is an important part of their work.

Given this review of past research on three important attributes, we next develop theory about individual perceptions of the legitimacy of EMNEs with particular attention to the entry of Chinese EMNEs into The Netherlands.

Theory Development

In this section, we expand on past research about individuals’ perceptions of legitimacy by including research on EMNEs from international business and business ethics. In so doing, we explain why legitimacy is fundamentally an ethical concept and how it applies to individuals in a host country evaluating EMNEs from another country, especially Dutch individuals evaluating Chinese EMNEs. We first present definitions of individuals and EMNEs. We then elaborate on the nature of legitimacy, its dimensions, its foundation in ethical theories, and the process by which individuals make legitimacy evaluations.

Individuals

Individuals are members of the general public residing in a host country, The Netherlands in our empirical study. They could be primary or secondary stakeholders (Clarkson 1995). In this way, our approach is person-centric rather than employee-centric, including persons inside and outside the organization (Gond et al. 2017, p. 226). We provide more details on how Dutch individuals evaluate the legitimacy of Chinese EMNEs below.

EMNEs

We define EMNEs “as international companies that originated from emerging markets and are engaged in outward FDI (foreign direct investment), where they exercise effective control and undertake value-adding activities in one or more foreign countries” (Luo and Tung 2007, p. 482). The number of EMNEs has expanded rapidly in the last decade (Luo and Tung 2018). Consistent with this, the amount of outward FDI from emerging markets has increased from 14.8% in 2006 to 41.2% in 2018.Footnote 1

Their rapid rise in recent decades has prompted inquiries about their ethical, social, and regulatory performance which is important for establishing legitimacy in host countries. All MNEs face a liability of foreignness when entering another country because MNEs lack experience in the host country; moreover, host country individuals, organizations, and governments lack experience with the entering MNEs and may be biased against them (Buckley and Casson 1976; Zaheer 1995). EMNEs, because of their under-developed home institutions (e.g., corruption and weak legislation), are often seen as having low accountability and weak business ethics (Luo and Tung 2007; Tang and Sun 2016). These disadvantages are known as the liability of emergingness (Madhok and Keyhani 2012).

These two liabilities pose significant challenges for EMNEs to be accepted in a host country, especially in developed markets. For example, Held and Bader (2018) found that German business students would prefer to work for an American company relative to a Chinese or Russian one. Han et al. (2018) found that inappropriate behavior, such as ignorance of sustainable development, a lack of respect toward the local culture, and hostile industrial relations, led to negative attitudes toward Chinese MNEs by the host government and the general public in European and African countries. Thus, gaining legitimacy is essential for EMNEs, yet research on what legitimacy of EMNEs actually is in host countries is very limited, especially from the perspective of individuals.

Legitimacy and Its Ethical Foundations

Legitimacy has displayed substantial elasticity that resulted in a plethora of definitions, measures, and theoretical propositions. While not all of them are fully compatible with one another, there is general consistency in most studies (Deephouse and Suchman 2008, p. 50). Meyer and Scott (1983, p. 201) provide one of first definitions:

A completely legitimate organization would be one about which no question could be raised. Every goal, mean, resource, and control system is necessary, specified, complete, and without alternative.

The most cited definition is from the seminal work of Suchman (1995, p. 574):

Legitimacy is a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions.

Although still in common use because of its generality and long-term impact, Suchman has refined his definition twice in subsequent reviews of recent literature (Deephouse et al. 2017; Deephouse and Suchman 2008). We adopt the definition presented in the latest review (Deephouse et al. 2017, p. 32) because it incorporates important elements from research re-considering the definition of legitimacy since Suchman (1995):

Organizational legitimacy is the perceived appropriateness of an organization to a social system in terms of rules, values, norms, and definitions.

The term appropriateness is included in the definitions of Suchman (1995), Zimmerman and Zeitz (2002), and Tost (2011). Moreover, this definition excludes the term “desirable” to avoid confusion with reputation (Bitektine 2011; Bundy and Pfarrer 2015; Deephouse and Suchman 2008; Foreman and Whetten 2012). Implicit in the term organization are the organization’s ends, means, and actions (Meyer and Scott 1983; Suchman 1995). Further, it recognizes the importance of the social system in which organizations and evaluators, especially individuals in our study, reside (Bitektine and Haack 2015; Scott 2014). Organizational legitimacy can be generalized to other subjects for which appropriateness to a social system is meaningful, such as practices, industries, and organizational forms (Elsbach 1994; Foreman and Whetten 2002; Jacqueminet and Durand 2020). We focus on an organizational form especially important in international business research, EMNEs.

As noted above, research on legitimacy from the perspective of individuals is expanding in the general management and organizational studies literatures, but this work remains disconnected from business ethics, a disconnect we find surprising given the overlap described below. We focus on examining how organizational legitimacy fits with the new pragmatism as an overarching theme and with deontological and utilitarian theories regarding particular dimensions. We start with an overview of business ethics from Wicks and Freeman (1998, p. 131):

Ethics in the context of business is about what the corporation strives to be (its goals or ends), how the corporation relates to its stakeholders, how managerial responsibilities are defined, and what ground rules will be used to limit and guide people’s behavior. … Taking the pragmatist approach involves seeing these matters as of primary importance, as “moral,” and as to a significant degree, up for grabs; thus, they need to be negotiated within the community of stakeholders who constitute a given corporation.

Legitimacy, commonly viewed in two ways, is clearly an ethical topic. The institutional view of legitimacy views organizations as constituted by norms, values, practices, rules, and languages of society comprised of communities of stakeholders, both geographic and virtual (Deephouse et al. 2017; Suchman 1995). The strategic view of legitimacy focuses on how managers seek to ensure their organizations are appropriate in society by negotiating with stakeholders to develop shared meanings of “what counts as good or bad organizing” (Pfeffer and Salancik 1978; Suchman 1995; Wicks and Freeman 1998, p. 132). Furthermore, “The domain of ethics spans both the ends and means of human (and corporate) activity. Ethics raises the issue of whether a given activity can be justified” (Wicks and Freeman 1998, p. 124). Assessments of legitimacy involve evaluating the ends and means of organizational activity (Deephouse and Suchman, 2008; Meyer and Scott 1983; Suchman 1995) and clearly fit within business ethics.

The entrance and growth of EMNEs into new host countries is fundamentally an equivocal phenomenon subject to a social process of interpretation within language communities (Wicks and Freeman 1998, following Weick 1979). Legitimation is the social process of negotiating and interpreting the appropriateness of organizational activity, EMNE expansion in our study. Legitimation in democratic societies like the Netherlands often involves the Habermasian concept of deliberative democracy involving ethical discourse and economic bargaining (Palazzo and Scherer 2006; Scherer and Palazzo 2007; cf. Suchman 1995). At a particular point in time, legitimacy is the current state of these interpretative efforts (Deephouse and Suchman 2008). Our research highlights a central component of this social process: Dutch individuals perceiving information about Chinese EMNEs and then assessing the propriety of them.

The Dimensions of Legitimacy and Their Ethical Foundations

The general concept of legitimacy as a summary judgment (Deephouse et al. 2017; Huy et al. 2014; Jacqueminet and Durand 2020; Meyer and Scott 1983; Suchman 1995; Tost 2011) is made more tractable by examining dimensions of the concept. One classification prominent in recent research on individuals’ perceptions is between validity and propriety. The latter represents the judgment by an individual of the appropriateness of an organization, type of organization, practice, or other organizational attribute; validity reflects the collective consensus that an organization is appropriate (Bitektine and Haack 2015; Tost 2011; Zelditch 2001). Evaluators in a social system, especially government officials and the media (Bitektine and Haack 2015), are important influences on propriety judgments by providing validity cues, an essential component in the legitimacy judgment process.

Research on legitimacy at individual, organizational, field, and societal levels has considered a variety of dimensions, also known as “types” (Bitektine 2011), “bases” (Scott 2014), or “criteria” (Deephouse et al. 2017). We specify four dimensions of legitimacy used in past research that are likely to be part of individuals’ perceptions of legitimacy of EMNEs entering new markets. Three propriety dimensions (i.e., regulatory, moral, and pragmatic legitimacy) overlap with the four categories of corporate social responsibility developed by Carroll (1979, 1991), and higher performance on these components should be reflected in higher legitimacy (Carroll and Shabana 2010; Rendtorff 2009, p. 155; Zheng et al. 2014). The fourth dimension, cultural-cognitive legitimacy, is a related to collective consensus on a number of components as part of the validity. It also overlaps with language communities valued in the new pragmatism and the pragmatist critical convergence (Visser 2019; Wicks and Freeman 1998).

The Propriety Dimensions of Legitimacy

We begin with dimensions associated with propriety, starting with regulatory legitimacy. It is gained by conforming to laws and rules set forth by governments and other regulatory authorities (Scott 1995), and it is similar with fulfilling the legal responsibilities of corporations (Carroll 1979). Individuals rated adherence to government regulations and other standards in what Elsbach (1994) called the prescriptive normativity dimension (this was published before Scott’s (1995) specification of regulatory legitimacy). EMNEs, like other firms, need to conform to local laws and regulations, but because EMNEs are from other countries, they are less familiar with local regulations than local firms are and thus face the liability of foreignness. Moreover, EMNEs may have less knowledge initially about laws and rules in developed countries than MNEs from other developed countries with similar political and economic systems (Jackson and Deeg 2008; Morgan et al. 2010), implying that they also need to overcome liability of emergingness (Gammeltoft et al. 2010; Madhok and Keyhani 2012). Chinese EMNEs recognize the importance of regulatory legitimacy; for instance, the Business Ethics policy of Huawei states: “We conduct business with integrity, observing international conventions and all applicable laws and regulations in the countries and regions where we operate. This is the cornerstone of operational compliance at Huawei, and has long been a core principle of our management team…” (Huawei Technologies Co., Ltd. 2019). From the perspective of individuals’ perceptions of EMNEs seeking to gain legitimacy, regulatory legitimacy reflects a person’s assessment that EMNEs are following the host country’s rules and laws.

The next dimension is moral legitimacy, which is gained by adhering to norms and values of society (Scott 1995; Suchman 1995). It overlaps with the ethical and discretionary responsibilities of corporations Carroll 1979). In addressing the changing nature of legitimacy in the context of globalization, Palazzo and Scherer (2006, p. 74) write: “moral legitimacy becomes the decisive source of societal acceptance for corporations in an increasing number of situations.” The moral legitimacy of Spanish Mutual Guarantee Societies was rated by individuals in Díez-Martín et al. (2013). As with regulatory legitimacy, EMNEs are less familiar with the norms and values of a host country and host country stakeholders are less familiar with EMNEs, so EMNEs need to overcome dual liabilities of foreignness and emergingness. Chinese EMNEs recognize the importance of moral legitimacy; for instance, “Alibaba Group is committed to creating a workplace that supports honesty, integrity, respect and trust” (Alibaba Group 2014, p. 1). From the perspective of individuals’ perceptions of EMNEs seeking to gain legitimacy, moral legitimacy reflects a person’s assessment that EMNEs are following the host country’s norms, values, and ethical principles.

The third dimension is pragmatic legitimacy (Suchman 1995). One source is satisfying stakeholders’ self-interests based on direct exchanges, consistent with the economic responsibilities of corporations (Carroll 1979). The pragmatic legitimacy of agricultural cooperatives is rated by individuals in Foreman and Whetten (2002). The ability of EMNEs to deliver satisfactory products and services is often debated in host countries simply because they are from emerging countries, consistent with the liability of emergingness (He and Zhang 2018; Wang et al. 2012). Nevertheless, some stakeholders welcome EMNEs because they are expected to bring capital and jobs to host countries. Pragmatic legitimacy also “can involve broader political, economic, or social interdependencies, in which organizational action nonetheless visibly affects the audience's well-being” (Suchman 1995, p. 578). Lenovo recognizes its importance in its Code of Conduct: “We are proud of Lenovo’s contributions to the quality of life and culture as well as the economic and social development of the communities where we do business” (Lenovo 2019, p. 47). From the perspective of individuals’ perceptions of EMNEs seeking to gain legitimacy, pragmatic legitimacy reflects a person’s assessment that EMNEs provide benefits to self and to society.

The propriety dimensions of legitimacy fit within the domain of business ethics and its concern for justification of business ends and means (Beauchamp and Bowie 2001; Wicks and Freeman 1998). Moral, regulatory, and pragmatic legitimacy are consistent with the focus of the new pragmatism in business ethics that highlights the extent to which organizations serve human purposes. Furthermore, pragmatic legitimacy invokes a utilitarian (consequentialist) ethos emphasizing outcomes of actions (Beauchamp and Bowie 2001; Chakrabarty and Bass 2015; Mill 1879). In our study, individuals assess how well Chinese EMNEs in The Netherlands benefit society, offer good products, boost the economy, etc. Similarly, moral legitimacy and regulatory legitimacy invoke a deontological ethos focusing on what is right and wrong (Beauchamp and Bowie 2001; Chakrabarty and Bass 2015; Kant 1871). In our study, individuals evaluate if they think Chinese EMNEs are following rules and norms of Dutch society.

The Validity Dimension of Legitimacy

We turn to the dimension associated with validity, cultural-cognitive legitimacy; it incorporates recognizability, comprehensibility, and cultural support (Scott 2014). Because EMNEs entering new markets seek to gain legitimacy, they are not yet ‘taken-for-granted,’ which is a common conceptualization of cognitive legitimacy closely related to institutionalization (Suchman 1995; Tost 2011). Furthermore, all MNEs face the challenge of pluralization resulting from the devaluation of tradition and individualization; this problematizes cultural homogeneity and renders taken-for-grantedness problematic, making management of MNEs and external assessment of them more complex (Palazzo and Scherer 2006). Ethical theorists also recognize that familiarity, knowledge, and common understandings need to be built in dialogic manner with stakeholder communities (Jensen and Sandström 2013; Rasche and Esser 2006; Visser 2019; Wicks and Freeman 1998), consistent with the process of legitimation presented above.

Past research takes a variety of approaches to cultural-cognitive legitimacy. Bitektine et al (2020 p. 112) ask: “Does the organization belong to any known category?” This is not relevant to a collective, that is, a group of EMNES from a particular country. Alexiou and Wiggins (2019) apply comprehensible and necessary or taken-for-granted as components. As noted above, EMNEs are not taken-for-granted, and the very act by a researcher in asking if a legitimacy subject is taken-for-granted makes it less so. Further, comprehensibility reflects an understanding of what is commonly understood, a second-order judgment (Haack and Sieweke 2020). Foreman and Whetten’s (2002 pp. 624, 632) claim to measure cognitive legitimacy of cooperatives as “viable and taken-for-granted,” but five of six items seem more like pragmatic legitimacy than cognitive legitimacy (e.g., “In general, I believe that co-ops are of little value to the small farmer.”). Humphreys and Latour (2013) take an alternate approach, measuring cognitive legitimacy as the inverse of response time in a categorization exercise. Haack and Sieweke’s (2020) commentary on Bitektine et al. (2020) highlights the challenge of measuring second-order validity dimensions. Tost’s (2011 pp. 692–693) meticulous theorizing excludes cognitive legitimacy “from the typology of the content of legitimacy judgments because (it) represents the absence of substantive judgment content.”

Given these difficulties, we take an alternate approach based on theorizing by Aldrich and Fiol (1994, p. 648), who view cognitive legitimacy for new ventures as the level of public knowledge about a new firm or industry and its activities. This approach was supported by Pollock and Rindova (2003) and Deeds et al. (2004) who found that volume of media reports as an indicator of public knowledge was associated with higher IPO valuations. It is also consistent with the recent empirical study of CSR implementation by subsidiaries of a large multinational by Jacqueminet and Durand (2020); they measured the validity conferred by corporate parent by the number of mentions of CSR in the corporate magazine. The level of public knowledge has been examined for different legitimacy subjects, such as downsizing (Lamertz and Baum 1998). We translate the level of public knowledge to the individual level. We recognize the role of public knowledge information in validating propriety decisions (Bitektine and Haack 2015; Tost 2011) and as an input to the comprehensibility component of cultural-cognitive legitimacy (Suchman 1995). A host of external regulators, rating agencies, social movements, social media, and traditional media evaluate the ends and means of organizations (Bromley and Powell 2012). However, individuals have limited information processing capabilities and filter information (Cyert and March 1963; March and Simon 1958). Thus, we define validating knowledge of an individual as the stock of validating information that is retained by an individual after filtering the large amount of validating information present at the collective level. Validating knowledge is used by an individual to make propriety judgments on particular dimensions of legitimacy (Bitektine and Haack 2015; Dornbusch and Scott 1975; Tost 2011; Zelditch 2001).

In the context of our study, EMNEs seek to overcome the liabilities of foreignness and emergingness by demonstrating to and communicating with host-country stakeholders that their characteristics and actions are recognizable and understandable (Scott 2014; Suchman 1995). EMNEs are relative newcomers—especially in advanced host countries—and individuals vary in their familiarity with them. Individuals form their perceptions of EMNEs based on whatever relevant information they perceive and then retain. This validating knowledge comes from several sources, not only government and traditional media (Bitektine 2011; Deephouse 1996) but also conversations with others and social media (Etter et al. 2018).

We conclude this theory section by enumerating other information that may influence Dutch individuals’ perceptions of Chinese EMNEs. One factor is industry; some industries like coal power face legitimacy challenges (Jahn et al. 2020). China’s country image may also be important, although country image, defined “as the total of all descriptive, inferential and informational beliefs one has about a particular country,” is a complex construct involving many dimensions (Martin and Eroglu 1993 p. 193).

Legitimacy is also influenced by the CSR behaviors and business ethics policies of Chinese EMNEs (De Blasio 2007; Du and Vieira 2012; Marano et al. 2017; Zhao 2012; Zheng et al. 2014), although these do not always compensate for fundamental ethical issues in core operations, sometimes known as greenwashing (Blanco et al. 2013; Scherer and Palazzo 2007). Relatedly, past scandals in China may have an influence. Wang and Laufer (2020) point out that Chinese social media is very active in raising attention to unethical corporate behavior, and some of these issues will be picked up by western media and media controlled by the Chinese government. For example, the 2008 scandal about melamine in Chinese infant formula powder that killed at least six babies and sent thousands to the hospital in China was picked up by the Chinese government, its media, and western media, including Dutch media (e.g., Moleman (2008) in Volkskrant). For example, The Economist not only reported at the time of the scandal but also in 2018 and 2019 (The Economist 2008, 2018, 2019). Similarly, Zheng et al. (2014 p. 409) reported numerous penalties for Chinese firms: “According to the China Securities Regulatory Commission (CSRC), from 2000 to 2010, 486 out of the 1,706 listed firms were penalized for having committed immoral behavior.” Some of this home country information gained from media may influence Dutch individuals. Direct experience as a customer, employee, supplier, or community member is of course important an important source. These multiple causes may influence individuals’ legitimacy assessments. However, our empirical research focuses only on creating measures for individuals’ legitimacy perceptions that have construct validity; only with valid measures can research examine possible causes in different settings.

To summarize our theory development, we have defined individuals, EMNEs, and legitimacy, including its dimensions and the process of gaining legitimacy (legitimation). We drew from ethical theories to demonstrate that legitimacy is fundamentally an ethical concept. We explained how individuals use validating knowledge to make propriety evaluations of organizations, with particular focus on our empirical setting: how Dutch individuals perceive the legitimacy of Chinese EMNEs in their country. Given this theoretical foundation, we next turn to how we developed a valid measurement scale.

The Scale Development Process

The empirical context of this study uses China as the home country and the Netherlands as the host country of individuals evaluating the legitimacy of Chinese EMNEs. This one-to-one setting (one home country and one host country) controls for cross-country variance, such as differences in individuals’ perceptions based on economic development, cultural heritage, and degree of ethnocentrism (Ahmed and d’Astous 2007). China is a relevant choice because that China is currently the largest contributor of global FDI outflow in the developing world. The share of outward FDI from China increased from 1.2% in 2007 to 12.6% in 2016, which makes China the second largest source of FDI in the world after the United States (UNCTAD 2017). We select The Netherlands as the host country for the following reasons. The Netherlands is a tolerant country (Smeekes et al. 2012), and the Dutch people are straightforward and direct when expressing different opinions on different topics. Second, the Netherlands is a valued location for MNEs because it provides access to the large European Union and has government policies conducive to FDI, including favorable tax treatment, leading this country of only 16.98 million people to be one of the largest FDI recipients in the world, ranking 5th in 2016 (UNCTAD 2017). Thus, Dutch people have relatively rich experience with MNEs that they can draw on for assessing the legitimacy of EMNEs.

Based on this context, we refine the definition of legitimacy for this study. The legitimacy of Chinese EMNEs in The Netherlands is the perceived appropriateness of Chinese EMNEs as a group in the context of the rules, values, norms, and definitions of Dutch society.

To measure individuals’ perceptions of this, we created a seven-step procedure based on the comprehensive procedures suggested by Churchill (1979) and Hinkin (1998). The steps are presented in Table 1. The first three steps generate items for survey and then reduces them using deductive reasoning and a pre-test. Steps 4, 5, and 7 address reliability, convergent validity, discriminant validity, and criterion-related validity; Step 6 investigates the relationships among the dimensions. Two samples were used in this procedure, following past research (Martin and Eroglu 1993; Sweeney and Soutar 2001; Davies et al. 2004). The first was a sample of students collected in step 2 for the pre-test. The second was a sample of Dutch citizens collected in step 4 used to test our scale among the population of interest.

Table 1 Procedure to develop a direct measurement scale
Full size table

Step 1: Item Generation

Hinkin (1998) summarized two approaches to create a list of preliminary items, deductive and inductive. A deductive approach is recommended when the theoretical foundation provides enough information to generate the initial set of items. It requires a clear theoretical definition of the construct under examination. An inductive approach is appropriate when the definition of the construct under examination is not yet well identified. With this approach, researchers usually develop scales inductively by asking a sample of respondents to provide descriptions of their feelings and understanding about the relevant concept. In our study, we use the deductive approach because the focal construct, legitimacy, is a well-researched concept. As noted above, existing literature provides a solid theoretical foundation for us to generate items. There is also a growing empirical corpus reviewed above that used survey questions to measure individuals’ perceptions of legitimacy mentioned above. These studies form the basis for using a deductive approach.

There were no existing items specifically for measuring the legitimacy of EMNEs. Following a deductive approach, we take the following sub-steps to generate initial items. We first conducted an extensive literature review. We started by searching with Google Scholar and EBSCO using the key words of legitimacy, legitimate, and legitimation. This resulted in over 500 journal papers, working papers, and books published during 1980–2017; the list is available from the first author on request. The publications reviewed are not only in management but also in other disciplines, such as political science, public administration, law, communication, accounting, sociology, and psychology. From this set, we found 49 empirical studies that quantified legitimacy, from which we found 133 original items used to measure legitimacy (available from first author). Of these, 32 use measures at the collective level, such as media coverage (e.g., Aerts and Cormier 2009; Bansal and Clelland 2004; Vergne 2011), population density (Baum and Oliver 1991; Rao 1994), numbers of organizations (Hannan and Carroll 1992; Carroll and Hannan 1989), certifications (Ivanova and Castellano 2012), and accreditations and memberships (Ruef and Scott 1998; Singh et al. 1986). We removed the collective indicators and ended up with 94 original items that measure individuals’ perceptions of legitimacy (available from first author).

Our second step was to filter these 94 items using the following criteria. First, we deleted 28 items that do not fit the definition of legitimacy used in this study but instead incorporate terms from other concepts like reputation or status. Two examples are: “The organization is viewed by business writers as one of the top firms in the industry” (Elsbach 1994, p. 87); and “Employees are proud to tell others they work at Bransford Sensors” (Certo and Hodge 2007, p. 470). Second, we deleted 17 items that measure internal legitimacy, the perception of organizational insiders. An example is: “You, your colleagues, your bosses: believe that all your organization’s actions help it achieve its goals” (Díez-Martín et al. 2013, p. 1959). Third, we deleted 24 items associated with a specific context which cannot translate to a group of organizations, EMNEs. An example is: “Because of the founder’s experience, the business has a founder who benefits the organization” (Pollack et al. 2012, p. 925). This step left 25 items.

In our third step, we combined items that appear to measure the same thing but in different ways into one item. For example, we translated and integrated the two items, “It seems to me that Suolo Inc. acts consistent with socially accepted norms and values” (Bachmann and Ingenhoff 2016) and “ZERTO Corporation complies with the norms and values of German society” (Jahn et al. 2020), into one item: “The Chinese firms conform to values held by our society” (Q1). This step ends up with 16 items.

In the fourth step, we compared these items to the definition and dimensions of legitimacy that we adopted to see if there were any substantial gaps. One item was added to reflect the terminology but not replicate the general definition: The business practice of Chinese firms is acceptable. We also added one item to measure pragmatic legitimacy considering the current economic situation in the host country, where unemployment was still higher in our survey years (2016–2017) than before the recession resulting from the financial and European debt crises of in 2007–2008. This item is: The Chinese firms provide opportunities to us to overcome the economic recession.

To assess face validity, we reviewed our items in detail with five experts: three professors in international business and two top managers who have rich international experience. All items passed their review. We also asked experts to provide new items that reflect the construct in context of Chinese firms in The Netherlands. Following experts’ suggestions, we added one item, I see a gap between formally agreed upon behavior and behavior in practice regarding rules and regulations. We ended up with 19 items for first data collection, measured on a Likert Scale from 1 to 5. Before the questionnaire was distributed, we did a pre-test with fifteen people by personal interviews to identify and eliminate potential problems. Based on their feedback we adjusted a few questions to make them easier to understand. The questions are presented in Appendix 1.

Step 2: The First Data Collection

The first survey was done in October and November 2016. We used an online survey instead of traditional methods such as letter, face-to-face, or telephone interview because online surveys are faster, cheaper, reduce participants’ burden, and may elicit greater variability in responses (Couper 2008). The survey was designed with the software provided by a survey company, Qualtrics. The online survey was pre-tested by a small group of people (four faculty members and six graduate students). Computers with different screen configurations and browsers and different smart phones with different systems were used in the pre-test to ensure consistent appearance of the survey. The questionnaire was slightly adjusted according to feedback from the pilot respondents.

The survey link was sent to the students of four programs (International MBA, Executive MBA, Modular MBA, and Part-time MSc) at Nyenrode Business Universiteit in The Netherlands. Because the majority of these programs are taught in English to classes that include international students and the English level of these students was sufficient for filling in the questionnaire, we used English in this survey. In the introduction of the survey, we indicated clearly that the questions were about Chinese companies in The Netherlands. There were 248 respondents; 39 were removed because of missing data, and another 8 respondents were removed because their response time was shorter than the item response time thresholds used by Wise and Kong (2005), specifically 20 s for 19 short items. This left 201 useable questionnaires. The average age is 34, and the percentage of females is 45%.

Step 3: Purify Items

The purpose of this step is to evaluate the number of items and explore the dimensions of the construct (Churchill 1979; Martin and Eroglu 1993). An exploratory factor analysis (EFA) with principal component procedures and varimax rotation was performed on the data collected in Step 2. An index of Kaiser’s measure of sampling adequacy (KMO = 0.789) and Bartlett’s test of sphericity (χ2 = 1230.88; df = 171; p < 0.001) suggested that the data are suitable for exploratory factor analysis (EFA). Eigenvalues greater than 1 were used to determine the number of factors. EFA extracts five factors accounting for 62.55% of the total variance. We used the following criteria to purify our list of items: (1) items should have communality higher than 0.4; (2) dominant loadings should be greater than 0.5; and (3) cross-loadings should be lower than 0.3 (Hair et al. 2006). Based on the results, we dropped Q8 because its factor loading is lower than 0.50 and Q6, Q7, and Q14 because of their high cross-loading. We dropped Q12 because it is the only item in its factor which is uninterpretable and may cause underidentification problems (Bollen 1989). In the end, 14 items and four factors remained, as presented in Table 2. We calculated Cronbach’s alpha to test the internal consistency of each factor. Table 2 shows that the alphas are greater than 0.6, which are acceptable in exploratory research (Hair et al. 2006, p. 778).

Table 2 Scale items and exploratory factor analysis result from initial sample
Full size table

The first factor includes three items: Q4 The business practice of Chinese firms is acceptable; Q5 The business practice of Chinese firms is proper; and Q13 The Chinese firms are good citizens. Because “appropriate is a covering term for both acceptable and proper” (Deephouse et al. 2017, p. 33), this factor is distinct from the other factors, and the other factors are closely linked to aforementioned dimensions of legitimacy, we follow Tost (2011) and refer to this factor as Generalized legitimacy.

The second factor includes three items: Q1 The Chinese firms conform to values held by our society; Q2 The Chinese firms are committed to meet norms and cultures standards that our community expects of foreign owned firms; and Q3 The Chinese firms conform to regulatory standards in our society. The three items were intended to measure moral and regulatory legitimacy, but they loaded on a single factor. That is, individuals do not perceive differences between moral and regulatory legitimacy that have been specified at the collective level (Deephouse et al. 2017; Scott 2014). Instead, this factor reflects theorizing that views these two dimensions as part of a superordinate dimension called sociopolitical legitimacy. According to Aldrich and Fiol (1994, p. 648), “Sociopolitical legitimation refers to the process by which key stakeholders, the general public, key opinion leaders, or government officials accept a venture as appropriate and right, given existing norms and laws.” The combined dimension reflects the attainment of acceptance by local stakeholders or conformity to socially constructed standards of behavior and legally established codes of conduct (Schultz et al. 2014). Some empirical research has found support for their hypotheses using Aldrich and Fiol’s (1994) conceptualization (e.g., Deeds et al. 2004; Wang and Qian 2011). Consistent with this work, we refer to this factor as Sociopolitical legitimacy.

The third factor includes three items: Q9 The Chinese firms are beneficial to our society; Q10 The Chinese firms provide opportunities to us to overcome the economic recession; and Q11 The Chinese firms provide good product/services to our society. The three items basically reflect audiences’ self-interest by measuring the extent to which people believe that the firms are beneficial to their society and themselves. Based on the explanation of pragmatic legitimacy provided by Suchman (1995), we refer to this factor as Pragmatic legitimacy.

The fourth factor includes five items: Q15 I follow news about Chinese firms; Q16 I discuss with friends and people around me about Chinese firms in Netherlands and/or other countries; Q17 I am aware that more and more Chinese firms come to Netherlands; Q18 How often do you hear or read that Chinese firms are being questioned or challenged in terms of their activities? and Q19 How often do you hear or read that Chinese firms are being endorsed in terms of their activities? These items measure individuals’ knowledge about focal firms. Consistent with our theorizing above, we refer to this factor as Validating knowledge.

Step 4: The Second Data Collection

Based on the results of step 3, we designed and conducted a second online survey which was conducted from December 2016 to February 2017. This second survey differs from the first one in terms of items, variety of respondents, approaches to access respondents, and language in use. However, we again stated in the introduction of the survey that the questions were about Chinese companies in The Netherlands. We started with the 14 items that remained after step 3. We added five additional items listed in Appendix 1 to assess criterion validity in step 7 and are discussed there.

Given our goal to measure Dutch individuals’ perceptions of the legitimacy of Chinese EMNEs, the second survey focuses on Dutch citizens and includes a diverse population in terms of education, profession, and age. Respondents were reached through the network of six recruiters as seeds. To ensure the sample is representative of the Dutch population (Hinkin 1998), the six recruiters varied on four criteria: location, age, education, and profession. The recruiters resided in the different regions of the country; their ages ranged from 22 to 55; their education levels ranged from vocational level to university level; and they were from different professions, including manager, social worker, teacher, and university student. The six seeds recruited respondents in two ways: (a) sending the survey link to their contacts, and requesting the contacts send it on; and (b) using Facebook to invite people in their region to fill in the questionnaire. Since the respondents were all Dutch citizens, the second questionnaire was in Dutch.Footnote 2

In total 480 people responded, from which 148 were removed because of missing data, and another 14 respondents were removed because their response time is shorter than the item response time thresholds used by Wise and Kong (2005), leaving 318 useable respondents. 44.3% of the respondents are male and 55.7% are female; the average age is 37, ranging from 18 to 71; and 28.0% of the respondents have university education. These sample statistics are similar to the Dutch population. According to Netherlands Statistics in 2016, 49.6% are male and 50.4% are female; the average age of the Dutch population is 41.2 years;Footnote 3 and 29.2% had university education.Footnote 4

Step 5: Reliability and Validity Assessment

In this step, we conducted confirmatory factor analysis (CFA) of data collected in the second survey to verify the four-factor model identified in previous EFA. We first assessed the overall fit of the model by using covariance-based SEM. The results in Table 3 show that the goodness-of-fit statistics (CFI = 0.756, SRMR = 0.098, RMSEA = 0.113) were not in line with the established criteria (CFI > 0.90, SRMR < 0.09, RMSEA < 0.08) (Hinkin 1998; Kim et al. 2012), necessitating further refinement of the model.

Table 3 Comparative analysis of models of various dimensionality
Full size table

After reviewing the results, we disaggregated the fourth factor, Validating knowledge. Q15, Q16, and Q17 measure the extent to which people know about or are aware of the firms and are akin to the amount of knowledge or volume of attention (Aldrich and Fiol 1994; Pollock and Rindova 2003). We label this factor Level of knowledge. Q18 and Q19 measure the valence of the individual’s knowledge in terms of endorsing and challenging (Deephouse 1996; Pollock and Rindova 2003). We labeled this factor Valence of knowledge. We also recognized that endorsing content and challenging content may have asymmetric effects on individual decision making (Barnett and Hoffman 2008; Baumeister et al. 2001; Haack et al. 2014; Suddaby et al. 2017). Because of this, we also split Valence of knowledge into its components to see if disaggregation better reflected individuals’ cognitive processing of validity information. Therefore, Q18 measures Challenging knowledge and Q19 measures Endorsing knowledge. We then analyzed the five-factor and six-factor models using CFA. The results presented in Table 3 indicate that the five-factor model and six-factor model both have acceptable fit. The six-factor model fits better than the five-factor model in terms of all fit statistics, and its superiority receives statistical support by testing the difference in the chi-squared statistics (63.21, df = 4, p < 0.0001). We also tested the overall fit of the models by using PLS-SEM (with ADANCO) and reach a similar conclusion. Therefore, we use the six-factor model in the following analyses.

We then assessed the reliability and validity of the six-factor model using PLS-SEM. To evaluate reliability, we apply two types of criteria. The first is related to each separate item, including item-to-item correlation and item-to-total correlation. We find two items (Q13 and Q17) have item-to-item correlation lower than the threshold of 0.3 (Hair et al. 2006, p. 137). We delete these two items to ensure sufficient internal consistency of the two relevant factors, Generalized legitimacy and Level of knowledge. With the remaining items, we assess composite reliability by Dijkstra–Henseler's rho (ρA) and Jöreskog's rho (ρc). Table 4 shows that each statistic for the six factors is higher than the 0.70 threshold (Henseler et al. 2016), indicating sufficient reliability.

Table 4 The assessment of composite reliability
Full size table

We next evaluated two types of validity, namely convergent validity and discriminant validity. Convergent validity is assessed by using two criteria, factor loadings exceeding 0.5 and the values of Fornell and Larcker’s Average Variance Extracted (AVE) exceeding 0.5, meaning that the latent variable explains more than half of its indicators’ variance (Hair et al. 2017). Table 5 shows that all factor loadings are above 0.5, and Table 6 shows that values of AVE are greater than 0.5. These results indicate a sufficient degree of convergent validity. To assess discriminant validity, we apply the commonly used Fornell–Larcker criterion (Hair et al. 2011), which is that the AVE of each latent construct should be greater than the highest squared correlation between a latent construct and any other latent construct. Table 6 indicates that AVE of the six factors are all higher than the squared correlations, implying the criteria are met.

Table 5 Factor loadings from second survey
Full size table
Table 6 The assessment of validity: Squared correlations and AVE
Full size table

Step 6: Inter-relationships among Dimensions

Based on the CFA in Step 5, we investigated the relationships among the dimensions. Applying our theorizing above and the literature on the formation of individual legitimacy judgments (Bitektine 2011; Ehrhart and Ziegert 2005; Tost 2011), we propose that the process goes through three stages: Validating knowledge; propriety judgments; and generalized judgment. Figure 1 depicts the proposed relationships.

Fig. 1
figure 1

The relationships among the factors. Note *, **. *** indicates that the relationship is supported at p < 0.05, p < 0.01, p < 0.001

Full size image

The three dimensions of validating knowledge are expected to affect the two propriety dimensions as follows. According to the exposure theory, repeated exposure to an object leads to increasingly positive evaluations of it (Zajonc 1968). This theory leads to the popular claim that “Familiarity leads to attraction” (Berscheid and Regan 2005, p. 177). Moreover, the environment processing metatheory predicts that the richness of the information processed by individuals influences their perception of the focal organization (Ehrhart and Ziegert 2005). Applying the theories to our research context, we expect that Level of knowledge is positively associated to Pragmatic legitimacy (relation ①) and Sociopolitical legitimacy (relation②).

We do acknowledge there is an opposite argument, “familiarity breeds contempt,” proposing that the more information individuals have about an object, the more likely they discover negative information (Norton et al. 2007, p. 97–98). To address this, we considered the effect of cognition in our specific context: Dutch individuals’ perception of Chinese firms. For many decades, China’s market system was not well developed. Chinese firms had the stigma of lagging technology, low quality, poor management, and strong links to government. However, in the past two decades, China made substantial improvements in economic development, technological advancement, and institutional quality (Ramamurti and Hillemann 2018; Zeng 2019). Furthermore, many successful Chinese multinational companies like Lenovo were emerging in the international stage. For example, in 2016, 110 Chinese companies were on the Fortune Global 500 list in 2016, compared to 46 in 2010 and 10 in 2000. Although some adverse cases generate publicity, the general trend is that Chinese firms are upgrading and acting according to international standards. It is likely that people will have better perceptions about Chinese firms if they receive more updated information about them and are more aware of their improvements. Therefore, the positive relationships between Level of knowledge on the one hand and Pragmatic legitimacy and Sociopolitical legitimacy on the other hand are defensible (relations ① and ②).

Expectancy theory (Vroom 1964) suggests that individuals who perceive an object to be favorable are more attracted to it. In our context, we expect that Challenging knowledge is negatively associated with Pragmatic legitimacy (relation ③) and Sociopolitical legitimacy (relation ④), and Endorsing knowledge is positively associated with Pragmatic legitimacy (relation ⑤) and Sociopolitical legitimacy (relation ⑥).

The last stage considers the individuals’ overall perception of the EMNEs, reflected by the factor Generalized legitimacy. Following Tost (2011), we expect this factor will be positively associated with the propriety judgments of Pragmatic legitimacy and Sociopolitical legitimacy (relations ⑦ and ⑧.)

Because this study uses survey data from the same respondents to test these relationships, there is the possibility of common method bias. To address this issue, we adopted approaches recommended by Chang et al. (2010), Podsakoff et al. (2003), and Podsakoff and Organ (1986). Specifically, to reduce the possibility of socially desirable responding and evaluation apprehension, we used an online survey. In the questionnaire administration, we assured anonymity and confidentiality of the responses. In addition, we used scales with different response formats, such as Likert scale for perception-based questions and numbers for fact-based questions.

Analysis and Results

We tested the predicted relationships using PLS-SEM on the data collected in the second survey. Following the suggestions of Henseler et al. (2009) and Hair et al. (2011), we tested the path coefficients and their significance with a bootstrapping procedure generating 5000 random samples. We examined the goodness of fit as suggested by Henseler et al. (2016). The results are presented in the first panel (A) of Table 7. The overall model estimation shows that the three criteria: SRMR, dULS, and dG are below the corresponding value of Hi99, indicating an acceptable fit.

Table 7 Results of the structural model estimation
Full size table

Table 8 presents the descriptive statistics and correlations for our major factors. Dutch individuals rate Chinese EMNEs near the midpoint of the 1 to 5 scale for the propriety dimensions (2.81 to 3.38). They are also relatively unfamiliar with them, with validating knowledge measures ranging between 2.05 and 2.35.

Table 8 Descriptive statistics and correlations
Full size table

The results for the paths are in the second panel (B) of Table 7, beginning with the influence of validating knowledge on propriety judgments. Level of knowledge is positively and significantly related to both Pragmatic legitimacy and Sociopolitical legitimacy (p < 0.05) in relations ① and ②. These results support our prediction that “familiarity leads to attraction” (Berscheid and Regan 2005) rather than “familiarity leads to contempt” (Norton et al. 2007). Challenging knowledge is negatively and significantly related to Pragmatic legitimacy and Sociopolitical legitimacy (p < 0.01, p < 0.0001), supporting relations ③ and ④. Endorsing knowledge is positively and significantly (p < 0.0001) related to Pragmatic legitimacy and Sociopolitical legitimacy, supporting relations ⑤ and ⑥. When comparing the magnitudes of the standardized coefficients from Endorsing knowledge and Challenging knowledge to Pragmatic legitimacy and Sociopolitical legitimacy, we see that the effect of Endorsing knowledge on Pragmatic legitimacy (β = 0.3254, p < 0.0001) is higher than the effect of Challenging knowledge on Pragmatic legitimacy (β = − 0.2345, p = 0.0016). In contrast, we see that the effect of the effect of Endorsing knowledge on Sociopolitical legitimacy (β = 0.3557, p < 0.0001) is lower than the effect of Challenging knowledge on Sociopolitical legitimacy (β = − 0.6086, p < 0.0001). One possible explanation of this finding is that the media is more critical when reporting on EMNEs’ (Chinese firms in this case) behavior related to conforming to the rules, norms, and values, and is relatively positive when discussing the contribution of the firms. The negative attitude toward Chinese firms regarding the sociopolitical legitimacy is likely sourced from the hot debates on the issues such as human rights or copyrights (Rapoza 2012; Xie and Page 2013). The positive attitude toward Chinese firms regarding pragmatic legitimacy is likely influenced by the fact that Chinese investment is helping Europe's economic recovery, such as creating jobs and generating tax revenue (Hanemann and Rosen 2012).

We next consider effect of the propriety judgments on Generalized legitimacy. Pragmatic legitimacy and Sociopolitical legitimacy are positively and significantly (p < 0.0001) associated with Generalized legitimacy, supporting relations ⑦ and ⑧. The standardized coefficients of the direct paths show that a one unit increase in the Sociopolitical legitimacy will increase Generalized legitimacy by 0.4995 units, and a one unit increase in the Pragmatic legitimacy will increase Generalized legitimacy by 0.2690 units. Thus the effect of Sociopolitical legitimacy is almost twice as large as the effect of Pragmatic legitimacy. For completeness, we tested a model of direct effects from Level of knowledge, Endorsing knowledge, and Challenging knowledge to Generalized legitimacy. None of the direct paths were significant.

The findings regarding the relationship between the factors are summarized in Fig. 1, which shows that all of the predicted relationships are supported.

Step 7. Assessment of Criterion-related Validity

We next examined criterion-related validity, specifically the influence of generalized legitimacy on Dutch citizen’s willingness to work for and purchase from Chinese EMNEs. These two outcomes have higher commitment and switching costs compared to the $100 crowdfunding experiment taken by an Amazon MTurk sample in Alexiou and Wiggins (2019) or the brand loyalty of University students to Montreal taxi firms in Bitektine et al. (2020). Taking a job is a very substantial commitment relative to the other criteria variables in past work using external stakeholders, not internal ones (cf. Jacqueminet and Durand 2020). The commitment associated with purchase varies with the characteristics of the item. Prominent Chinese EMNEs in The Netherlands offer a number of high commitment, high switching cost items such as mobile phones (Huawei and Xiaomi), appliances (Haier), and automobiles (Byd), including the electric car segment that is expanding with societal concern about global warming (SAIC Motor).

The effects of legitimacy on decisions to work and purchase are supported by much management research, including resource-dependence (Pfeffer and Salancik 1978), new institutional theory (Suchman 1995), and entrepreneurship (Shepherd and Zacharakis 2003). Legitimacy is considered a resource that enables organizations to access other resources needed to survive and grow (Pfeffer and Salancik 1978; Zimmerman and Zeitz 2002). Stakeholders are more likely to supply these resources to organizations that are legitimate (Suchman 1995). As Brown (1998, p. 349) stated: “legitimate status is a sine qua non for easy access to resources, unrestricted access to markets.” Two important resources provided by individuals are labor services and product purchases which are influenced by willingness to work for and purchase from Chinese EMNEs. Williamson (2000, p. 30) predicts that “legitimacy of an organization will be positively related to an organization's ability to successfully recruit potential job applicants.” Both management and marketing research shows that legitimacy is important to customers (e.g., Handelman and Arnold 1999; Shepherd and Zacharakis 2003).

We assessed criterion-related validity by testing with PLS-SEM the model used in Step 6 with two additional paths: one is from Generalized legitimacy to Willingness to work, and the other is from Generalized legitimacy to Willingness to buy. We used four items from the second survey of Dutch citizens to measure Willingness to work (Zhang et al. 2020) and one item to measure Willingness to buy (Laroche et al. 2005); the items can be found in Appendix 1. A single-item measure is equally valid as multiple-item measures for simple constructs like willingness to buy (Bergkvist and Rossiter 2009). The overall model estimation shows that the three criteria, SRMR, dULS, and dG (0.0522, 0.4176, and 0.2154) are all below the corresponding value of Hi99 (0.0658, 0.6631, 0.7149), respectively, indicating an acceptable fit. The path coefficient from Generalized legitimacy to Willingness to work is significant (β = 0.3581; p < 0.0001). The path from Generalized legitimacy to Willingness to buy is significant (β = 0.1786; p = 0.0142). These results support the relationships found in the literature, providing criterion validity. Together with our results in step 6, there is evidence of construct validity for our measurement scale.

Discussion

EMNEs are a type of organization becoming increasingly important globally. When EMNEs enter a new host country, stakeholders evaluate if their ends and means are good for the country in terms of its norms, values, ethical principles, laws, and regulations, often using dialogic processes inherent in deliberative democracy and the pragmatist critical convergence (Brønn and Vidaver-Cohen 2009; Kostova and Zaheer 1999; Palazzo and Scherer 2006; Suchman 1995; Visser 2019; Wicks and Freeman 1998). To gain legitimacy, EMNEs entering an advanced country must surmount the dual liabilities of foreignness and emergingness (Held and Bader 2018; Madhok and Keyhani 2012; Zaheer 1995). Recent research has drawn attention to individuals’ perceptions of legitimacy (Bitektine and Haack 2015; Deephouse et al. 2017; Suddaby et al. 2017; Tost 2011). Based on prior studies (e.g., Churchill 1979; Hinkin 1998), we used a seven-step procedure to develop and validate a scale to measure individuals’ perceptions of legitimacy of Chinese EMNEs entering The Netherlands. We offer the following contributions to the literature and suggestions for future research.

Research Implications

Our first contribution is theoretical. While there has been much research on legitimacy, very little has used ethical theories to examine the concept. We put an ethical foundation under legitimacy and its dimensions, using the new pragmatism, deonotology, and utilitarianism, and demonstrate that legitimacy is fundamentally an ethical concept. Future research could apply ethical theories in more depth to the study of legitimacy. For example, ethical principles centered on the natural world as a stakeholder may reframe legitimacy in the context of the developing climate crisis (Driscoll et al. 2012; Starik 1995). Another alternative especially appropriate for EMNEs entering host countries is the Heideggerian principle of dwelling (Ladkin 2006). A third theory, and appropriate for settings involving Chinese, is Confucianism, which has already been applied to shareholder appropriation (Du 2015) and business ethics (Ip 2009). These efforts also would demonstrate the value of business ethics to other theoretical domains, such as entrepreneurship (Aldrich and Fiol 1994), institutional theory (Deephouse et al. 2017), and international business (Kostova and Zaheer 1999).

Our second contribution is theoretically developing and finding empirical support for a model of the legitimacy judgment process of individuals. Our central innovation was developing the concept of validating knowledge held by individuals. Past research highlighted the presence of validity cues provided by validating institutions that are used as inputs into propriety evaluations (Bitektine and Haack 2015; Tost 2011). We improve on this work by incorporating the fact that individuals variably attend to and filter these cues (March and Simon 1958) resulting in a stock of validating knowledge. This concept had three components: level of knowledge, endorsing knowledge, and challenging knowledge. Our specification and testing of the nature of validating knowledge held by individuals for use in propriety judgments provides a micro-foundation under research using media and discourse at the collective level.

Our third contribution concerns the dimensions of legitimacy inherent in individual perceptions of the legitimacy of EMNEs. We found six dimensions: generalized legitimacy, sociopolitical legitimacy, pragmatic legitimacy, level of knowledge, endorsing knowledge, and challenging knowledge. Our finding of a separate factor for generalized legitimacy is consistent with work by Meyer and Scott (1983), Suchman (1995), Tost (2011), Deephouse et al. (2017), and Jacqueminet and Durand (2020) but not part of the work by Alexiou and Wiggins (2019) or Bitektine et al. (2020).

We found individuals viewed regulatory and moral legitimacy virtually identically, with items for these two dimensions loading on a single factor which we called sociopolitical legitimacy. In other words, individuals process together validating information about compliance with formal rules and informal norms from different field-level validating institutions, such as like media, government regulators, and judiciary (Bitektine and Haack 2015). This finding is contrary to the dimensions specified by Scott (1995) and Deephouse et al. (2017) and identified in field-level research (e.g., Deephouse 1996). It is consistent with Aldrich and Fiol’s (1994) definition in entrepreneurship theory, positive connectedness of laws and moral values (Bascle 2016), deontological reasoning addressing questions of right and wrong (Beauchamp and Bowie 2001; Chakrabarty and Bass 2015), and results supporting this dimension in Bitektine et al. (2020). Although crisis managers and crisis management researchers recognize that media reports and regulatory action do not happen at precisely the same time, individuals making legitimacy assessments may see them as commonly occurring together. Future research should be alert to the potential for individuals to treat regulatory and moral legitimacy together yet also investigate contingencies that may separate them.

We found that individuals use endorsing knowledge and challenging knowledge differently not only in terms of direction (positive vs. negative), but also in terms of magnitude. Specifically, endorsing knowledge is more relevant to Chinese EMNEs’ pragmatic legitimacy, while challenging knowledge is more relevant to their sociopolitical legitimacy. These have been viewed as components of holistic measures of media legitimacy at the collective level (Bansal and Clelland 2004; Deephouse 1996; Vergne 2011). Therefore, future research on individual perceptions of legitimacy or other ethically based assessments of the goodness of corporate behavior may find it better to disaggregate endorsing and challenging knowledge (Baumeister et al. 2001; Dean 2004; Haack et al. 2014; Westphal and Deephouse 2011). Future research could also explore how individuals process endorsing information from one validating institution and challenging information from another one to examine if some sources are more authoritative or if bad news is stronger than good (Baumeister et al. 2001).

Future research could compare the effects of individual perceptions of legitimacy and of CSR on willingness to work and purchase. The relationship between companies’ CSR involvement and employer attractiveness goes back at least 20 years to a landmark paper by Turban and Greening (1997) and includes closely related concepts of corporate social performance and corporate citizenship (Backhaus et al. 2002; Kim and Park 2011; Lin et al. 2012). Research in marketing on the influence of CSR on customer loyalty is similarly long-lived (Sen and Bhattacharya 2001; Marin and Ruiz 2007; Pérez and del Bosque 2015). This growing body of work is now known as micro-CSR, and recent reviews report conflicting findings for the effects of CSR on employees, potential employees, and customers (Gond et al. 2017; Rupp and Mallory 2015). One possible reason is the many disconnects between CSR and corporate action consistent with greenwashing (Blanco et al. 2013; Laufer 2003).

Many companies engage in substantial philanthropy but do not touch upon the ethical challenges of their key operations. Others engage in self-regulation but do not include external, especially critical, voices in the development of rules or the monitoring of success. Companies sometimes position themselves as sustainable and drown the readers of their CSR reports in technical data but do no more than comply with basic environmental laws. (Scherer and Palazzo 2007, p. 1114).

Another reason is that micro-CSR studies focus on CSR behaviors. In contrast, legitimacy encompass all the ends and means of a corporation (Wicks and Freeman, 1998), including its economic, legal, ethical, and discretionary social responsibilities (Carroll 1979). Thus, individuals’ perceptions of legitimacy should be more accurate predictors of employee engagement and customer loyalty than perceptions of CSR.

Our study of individuals as the micro-foundations of legitimacy warrants comment on how individual perceptions influence legitimacy at the collective level. One way this occurs is through political processes associated with deliberative democracy (Palazzo and Scherer 2006; Scherer and Palazzo 2007). Elected members of government generally stay in touch with their constituents, especially in well-functioning democracies like The Netherlands. Moreover, individuals may communicate by social media, traditional media, or talking with friends which minimize the gap between individual-level legitimacy and collective indicators such as media coverage. In addition, with corruption relatively low in the Netherlands, it is unlikely that Chinese firms gain legitimating endorsements from ministries, associations or media through secret operations (i.e., bribery). This minimizes the difference between individual-level legitimacy and collective indicators such as media coverage, certifications, accreditations, and memberships. Thus, we expect there to be a close relationship between individual perceptions and collective indicators of legitimacy in The Netherlands.

Taken together, this study provides a foundation for future research of individuals’ perceptions of legitimacy. We studied a specific subject, EMNEs from China, within a particular social system, The Netherlands, and this design controlled for a variety of confounding factors. However, we believe our scale is adaptable to both Emerging and Developed Market MNEs in other countries. This could be done primarily by altering the home country in the survey items and host country where the survey is administered. Secondary modifications would be necessary for local contingencies (cf. Jahn et al. 2020). For example, the item regarding the economic recession (Q10) is context-specific and may not apply in non-recessionary contexts. However, this reminds us that economic outcomes are important for pragmatic legitimacy, so researchers should create a new item that captures this element.

Future research could use these validated measures to examine the determinants of EMNE legitimacy. These include industry (Jahn et al. 2020), home country image (Martin and Eroglu 1993), CSR (Panwar et al. 2014), and business ethics (Zheng et al. 2014), where CSR represents behaviors toward stakeholders and business ethics represents corporate norms and values (Muller and Kolk 2010; Zheng et al. 2014). Future research could also use these measures to assess legitimacy of EMNEs and developed market MNEs from different countries, such as Indian, American, and German MNEs in The Netherlands, to examine the relationships explicated in Bascle (2016) concerning the dynamic legitimacy for MNEs with different levels of accumulated legitimacy. This research could also be used as part of the ongoing process to develop better measures of individual perceptions of organizational legitimacy (Flake et al. 2017; Haack and Sieweke 2020).

Managerial Implications

This study provides practical implications as well. First, our study shows that while legitimacy exists in individuals’ minds as a generalized evaluative judgment on the appropriateness of corporate practices, generalized legitimacy is determined directly by sociopolitical legitimacy and pragmatic legitimacy. Therefore, firms need to behave in a socially responsible and ethical manner to gain generalized legitimacy, including conforming to instrumental demands to gain pragmatic legitimacy and conforming to rules, norms, and ethical principles to gain sociopolitical legitimacy. Firms also need to influence what stakeholders know about them through advertising and social media promotion to enhance knowledge about their actions. For example, substantive CSR actions should enhance pragmatic and sociopolitical legitimacy, and strategic communication of these actions should enhance individuals’ validating knowledge (Bachmann and Ingenhoff 2016; Du and Vieira 2012).

Second, the empirical results show that sociopolitical legitimacy contributes to generalized legitimacy more than pragmatic legitimacy does. Thus, Dutch citizens value more conformity by Chinese EMNEs to rules and norms of Dutch society than the perceived benefits provided by them. Our data also show that Dutch individuals on average rated Chinese EMNEs’ sociopolitical legitimacy (2.81) lower than the pragmatic legitimacy (3.38). This combination implies that the compensation effect between the two legitimacy dimensions is limited. Due to the relative low effect of pragmatic legitimacy, the positive effect of a high pragmatic legitimacy is not strong enough to offset the negative effect of a low sociopolitical legitimacy. This finding suggests that Chinese EMNEs with scarce resources should focus on improving sociopolitical legitimacy, such as using values-focused CSR and improving ethics policies (Zheng et al. 2014).

Finally, the results show that generalized legitimacy enhances access to resources and markets in the context of EMNEs in an advanced country. This suggests that the importance of building legitimacy goes beyond survival—it also is imperative for growth and sustainability. Managers of Chinese EMNEs should put enough efforts in choosing and implementing programs to enhance legitimacy, such as CSR, as long as the programs are genuine (Blanco et al. 2013; de Roeck and Delobbe 2012).

Limitations

Despite the theoretical contributions and practical implications, this study has limitations that warrant future research. First, this study only focuses on the legitimacy perceived by individuals in a society. It would be valuable to measure the perceptions of specific stakeholders and compare the differences among them. Different stakeholders may have different criteria regarding legitimacy, and some dimensions may be more relevant to one group than others (Bascle 2016; Deephouse et al. 2017; Finch et al. 2015; Haack and Sieweke 2018). Moreover, we did not measure the nature of individuals’ stakeholder relationships with Chinese EMNEs; future surveys could do so.

Second, our empirical study is focused on Chinese firms operating in the Netherlands, which limits the direct application of the result to other EMNEs from other emerging countries in other host countries. As noted above, extending the research to other emerging countries and other host countries would require only minimal changes to the items to examine external validity.

Third, the relative importance to individuals of different items measuring legitimacy was not measured (Bascle 2016); this limitation pertains to many other studies of individual perceptions of legitimacy (e.g., Alexiou and Wiggins 2019; Bitektine et al. 2020). For instance, some individuals may value environmental performance more than gender diversity or fit within a particular category (Bitektine et al. 2020; Jacqueminet and Durand 2020). Such weights may be important in determining overall legitimacy judgments and subsequent actions (Bascle 2016), such as the purchase and employment decisions that we studied or the implementation of CSR practices within a large multinational (Jacqueminet and Durand 2020). Moreover, weights can change over time or place (Bascle 2016; Jacqueminet and Durand 2020), so that environmental CSR may be more important in one period or country and health and safety in another period or country. Relative importance of an item or dimension may be influenced by six parameters that influence the importance of underlying norm; future research should examine the role of calculability, connectedness, consensus, power heterogeneity, prescriptivity or proscriptivity, and role overlap (Bascle 2016). Relative importance may also inform the study of compensating effects among items and dimensions (Bascle 2016). Research using set theoretic methods may be useful to assess compensation effects on legitimacy (Jacqueminet and Durand 2020; McKnight and Zietsma 2018).

Fourth, this study does not look into specific industries, and the importance and criteria of legitimacy are different across industries. For example, firms from a new industry may need to put more effort in building legitimacy than those in well-known industries (Aldrich and Fiol 1994), and polluting industries have greater legitimacy concerns than high tech (de Roeck and Delobbe 2012; Jahn et al. 2020). Future studies should pay attention to the linkage between the features of industry and dimensions of legitimacy and develop a theoretical framework for firms to choose appropriate legitimacy strategy in different host countries and industries.

Conclusion

This study contributes to the growing research program to study individuals’ perceptions of legitimacy (Alexiou and Wiggins 2019; Bitektine et al. 2020; Finch et al. 2015; Haack and Sieweke 2018; Jahn et al. 2020; Kim et al. 2014; Panwar et al. 2014; Yang et al. 2012). We contribute to this ongoing research effort (Flake et al. 2017; Haack and Sieweke 2020) by focusing on the legitimacy of EMNEs, an organizational form that requires legitimacy to overcome the liabilities of foreignness and emergingness (Buckley and Casson 1976; Held and Bader 2018; Madhok and Keyhani 2012; Zaheer 1995). We deduced that evaluations of legitimacy are fundamentally ethical analyses made by stakeholders about the goodness of organizational ends and means (Beauchamp and Bowie 2001; Wicks and Freeman 1998). The legitimacy judgment process model and measurement scale that we developed and validated could applied in a variety of settings to examine the legitimacy of other organizations, its antecedents, and consequences.