[The] temporal and contextual factors set the boundaries of generalizability, and as such constitute the range of the theory. Scholars who study the effects of time and context on people and events keep asking nagging questions like, Would your predictions hold in Japan, with a blue collar population, or across time periods? Unfortunately few theorists explicitly focus on the contextual limits of their propositions. In the efforts to understand a social phenomenon they tend to consider it only in familiar surroundings and at one point in time.
Whetten (1989: 492)
The challenge of integrating context when theorizing is not unique to research on management in emerging economies (MEE). Principally, all management scholars aspiring a theoretical contribution should be concerned about context, as emphasized by Whetten (1989: 492) in his oft-cited guidelines.
However, insufficient attention to contextual boundary conditions continues to inhibit management research as highlighted in recent review papers in this journal (Jia, You, & Du, Reference Jia, You and Du2012; Zhang, Chen, Chan, & Ang, Reference Zhang, Chen, Chen and Ang2014). MEE research offers opportunities to build more robust theoretical foundations for management scholarship, and to generate insights relevant to practice. Therefore, the assumption that a theory is ‘universally valid unless proven otherwise’, is normally not a good starting point for papers submitted to MOR. Rather, we expect authors to intellectually engage with the contextual boundaries of their theorizing and with context-specific issues that may impact their empirical findings.
Context can be incorporated in a variety of different ways. The most appropriate approach depends on both the research question and the empirical context. This editorial provides suggestions regarding how consideration of empirical contexts can advance management research. Deep contextualization can be a particularly powerful approach to generate novel theoretical insights; it is not the only way to engage with context. In addition, I recommend stating assumptions explicitly, designing insightful comparisons, using multi-context settings, and reflecting deeply on context-specific and general aspects of the research.
THEORY GROUNDED IN CONTEXT
Chinese management research in particular draws on a long tradition of studies that investigate the specific nature of the Chinese economy, and on that basis generate influential theoretical advances (Boisot & Child, Reference Boisot and Child1996; Nee, Reference Nee1992; Redding, Reference Redding1993). For example, studies of guanxi have enriched our understanding of management in the Chinese context (Chen, Chen, & Huang, Reference Chen, Chen and Huang2013; Liu, Keller, & Hong, Reference Liu, Keller and Hong2015). Although concepts like guanxi ties, guanxi quality, and guanxi practices correspond to concepts in the mainstream networks, social capital, and relationship marketing literatures, those literatures cannot satisfactorily explain phenomena in Greater China (Chen et al., Reference Chen, Chen and Huang2013), while China-based research highlights differences in networking across societies and groups.
A key challenge for studies of such novel phenomena is the integration with international scholarly discourses. Principally, this can be done in two ways, by using existing theories to explain empirical puzzles (‘contribution of theory’), or by modifying theories through the study of novel phenomena with the aim to enhance the theory (‘contribution to theory’, Whetten, 2007). A good practice example that enhances theory by developing new constructs from an in-depth investigation of a non-conventional context is Mair, Marti, and Ventresca's (Reference Mair, Marti and Ventresca2012) analysis of bottom of the pyramid markets in Bangladesh. They contribute to the institutional voids literature by identifying two sets of activities, redefining market architecture, and legitimating new actors that are critical for building inclusive markets.
Another approach is to start from existing theory, but focus on a context where the concepts and relationships as stated in the theory do not appear to apply. Such a grounded approach can extend the theory by providing novel insights into contextual boundaries or contextual moderators. For example, Cui, Liang, and Lu (Reference Cui, Liang and Lu2015) test hypotheses from the literature in a distinctly different context, and use their contrarian results to discuss contextual boundaries. As another example, Guo and Giacombe-Miller (Reference Guo and Giacobbe-Miller2015) inductively develop a concept previously developed in a US-based study, and identify similarities and differences in the dimensions of the construct.
Make Assumptions Explicit
Good theory is built on parsimonious assumptions that simplify the complexity of the real world. Clearly stated assumptions establish the building blocks for a theoretical argument, and thereby set the boundary conditions for a theoretical concept or model.
For example, principal agent theory applications to corporate governance usually assume that companies have (a) principals (i.e., shareholders) who are shareholder-wealth oriented, and (b) agents, or managers, whose objectives include personal wealth but also prestige and ease of life. With these assumptions, deviations from shareholder wealth provide a proxy for the extent of agency conflicts. These assumptions are a sensible approximation for US-listed companies owned by financial investors, and hence common in corporate governance research. However, where these assumptions are not close approximations to reality, then the empirical test is grounded in spurious assertions, and thus becomes useless as a test of the original hypotheses. Specifically, testing agency theory on Chinese listed firms has to reflect the fact that about half of these firms have substantial state ownership, often majority ownership (Delios, Wu, & Zhou, Reference Delios, Wu and Zhou2005; Morck, Yeung, & Zhao, Reference Morck, Yeung and Zhao2008). At the same time, top managers often follow career paths within the state sector (Brødsgaard, Reference Brødsgaard2012). Hence, neither of the two assumptions holds. A rigorous application of agency theory requires first establishing the objective functions of the principals and of the agents, and on that basis to design an appropriate empirical test (Estrin & Prevezer, Reference Estrin and Prevezer2011; Nee, Opper, & Wong, Reference Nee, Opper and Wong2007).
A set of assumptions may also establish a specific constellation of home and host societies in which to study international business. For example, Child and Marinova (Reference Child and Marinova2014) define the object of their study as MNEs originating from a politically stable emerging economy investing in developing countries with weaker institutions. Under these boundary conditions, they propose that such MNEs rely more on their political connections and develop their strategies on the back of government-to-government relations. Similarly, Li, Newenham-Kahindi, Shapiro, and Chen (Reference Li, Eden and Beamish2013) explore MNE to host government bargaining for the case of Chinese firms, and develop a model that shows distinctly different lines of causality than the established model in the literature, notably a different role of the home country government. Both papers elicit important insights for explaining the global economy, without invoking an assumption that their model would apply to all forms of government.
In setting the theoretical boundary conditions, general statements such as ‘an industry with characteristics x, y, z’ normally provide a more solid foundation for a theoretical discussion than descriptions such as ‘the Chinese brewing industry’. For example, authors may state at the outset that their theoretical arguments apply to industries characterized by a) rapid consolidation in volatile markets, b) distinct local markets due to high barriers to inter-regional trade. The empirical analysis then can be conducted using data from the brewing industry. In the discussion section, authors can then reflect in which other industries (e.g., other FMCG with high transportation costs) and geographies (e.g., countries like India in the near future, or regions such as Eastern Europe in the 1990s) similar conditions may be found. Further, they may discuss each of the assumptions: how would the results change if the study was conducted a) in a mature market with stable industry structure? b) a FMCG industry with low barriers to trade due to lower transportation costs? In this way, clarity in the assumptions leads to a structured and thus informative discussion of how and why context does (or does not) matter for the study.
Design Insightful Comparisons
Strong insights on contextual boundary conditions can be derived from comparisons across different, carefully-chosen contexts. If the objective of a study is to demonstrate the general validity of a theory or hypothesis, then an empirical test needs to be replicated in a very different context (Tsang & Kwan, Reference Tsang and Kwan1999). Likewise, if the purpose of a study is to show the existence of differences in a phenomenon between countries, or the magnitude of such differences, then choosing countries with large context difference is most likely to find evidence in favor of differences in phenomena (see for examples Liu et al., Reference Liu, Keller and Hong2015; Zhang, Liu, & Liu, Reference Zhang, Liu and Liu2015).
If, on the other hand, the purpose of the comparison is to identify how context variables influence a phenomenon, then contexts that vary by only a small number of critical parameters allows more concise inferences about the sources of observed variations. Comparisons between the US and China, while popular, rarely meet this criterion because the two countries vary on a large number of dimensions (White, 2002). Three types of comparative studies are likely to be particularly helpful.
First, comparisons within Greater China enable fine-grained analysis of contextual influences. Greater China provides an interesting laboratory because businesses within this region share some traditions and cultural values. Yet, their contexts differ with respect to business-to-government relationships, ownership structures, and international openness, for example. This variation enables comparative studies to generate new insights into how and why, for example, institutions impact on environmental management practices (Child & Tsai, Reference Child and Tsai2005), leadership styles (Farh, Earley, & Lin, 1997), guanxi practices (Chen et al., Reference Chen, Chen and Huang2013; Luk, Yau, Sin, Tse, Chow, & Lee, Reference Luk, Yau, Sin, Tse, Chow and Lee2008) and alliance partner selection (Ahlstrom, Levitas, Hitt, Dacin, & Zhu, Reference Ahlstrom, Levitas, Hitt, Dacin and Zhu2014).
Second, comparisons of China with other emerging economies are important to gain a better understanding to what extent the Chinese context is unique rather than typical for an emerging economy. Despite repeated calls for more comparative research between emerging economies (Meyer, Reference Meyer2006; White, 2002) such research involving China remains scarce, apart from comparative studies of China and Russia with respect to networking (Batjargal, Reference Batjargal2007; Michailova & Worm, Reference Michailova and Worm2003) and knowledge sharing practices (Michailova & Hutchings, Reference Michailova and Hutchings2006). Also, historical matching of countries facing similar challenges at different points in their history can be insightful, such Kang, Jiang, Kang, and Ke's (2009) comparative analysis of early stage foreign investment from respective Japan and China.
Third, comparing regions within countries offers opportunities for more fine-grained analysis of theoretical ideas arising in cross-national research. Some large countries, such as China, India, and even Vietnam, have distinct sub-national entities that vary on critical variables such as institutional development. Such variations offer great opportunities to examine, for example, how sub-national institutional variations impact on the attraction of foreign investors and the growth of private firms (Li, Eden, & Beamish Reference Li, Eden and Beamish2013; Nguyen, Le, & Bryant, Reference Nguyen, Le and Bryant2013). However, few studies analyze how variations across regions and provinces in emerging economies influence management practices.
Design Multi-country Studies
Studies using data from a large number of countries allow empirical testing of the impact of contextual variables. First, direct effects of national-level variables capture variations in the intensity of a phenomenon or behavior, but are usually of least theoretical interest. Second, moderating effects capture variations in the intensity of an effect due to variations in contextual variables such as cultural norms or institutional development. Recent studies employ HLM methodologies to moreover analyze variations of effects across contexts (Jia et al., Reference Jia, You and Du2012).
However, equivalent data from diverse countries are rarely available to management researchers because their constructs tend to be abstract. One interesting approach to overcome this constraint is meta-analysis, a technique that uses empirical findings from earlier studies as data. The contexts of those earlier studies can be introduced as explanatory variables. For example, Luo, Huang, and Wang (Reference Luo, Huang and Wang2012) used a meta-analysis of the guanxi literature to examine variations in findings between, among other tests, mainland and overseas Chinese and found guanxi to have stronger performance impact in mainland organizations. With a larger set of countries, country-level variables may be introduced in a meta regression analysis; for example, Carney, Gedajlovic, Heugens, Van Essen, and Van Oosterhout (2012) examine the effect of business group affiliation on firm performance across national contexts varying in their institutional development.
Methodological innovations open new opportunities to exploit multi-country datasets. For example, Judge, Fainshmidt, and Brown (Reference Judge, Fainschmidt and Brown2014) employ ‘fuzzy-set qualitative comparative analysis’ to identify national institutional configurations associated with wealth and equality. Analyzing qualitative evidence from the four BRIC countries, Puffer, McCarthy, Jaeger, and Dunlap (Reference Puffer, McCarthy, Jaeger and Dunlap2013) investigate favors-related practices and identify communalities across the four BRIC countries, and then develop their theory on the basis of the identified communalities. Such new methodologies of analyzing qualitative and quantitative data from multiple countries may thus help developing new theoretical insights into how contextual conditions matter.
Discuss Boundary Conditions
Theoretical contributions require a critical reflection over the boundary conditions of both the theoretical arguments and the empirical evidence. The discussion section is usually the appropriate place for such reflections. A single sentence ‘we suggest testing our hypotheses with data from another country’ is unlikely to stimulate future theoretical work. For example, I notice a tendency to generalize from China-based research by implying that theories and models developed for and tested in mainland China apply to other emerging economies. This assumption is no better than the wide-spread practice of deriving ‘universal’ knowledge from US data alone (Henrich, Heine, & Norenzayan, Reference Henrich, Heine and Norenzayan2010). China has much communality with other emerging economies, but also very distinct features. Understanding these communalities and distinctions is critical to advance MEE research.
For scholars working within a single context, such as China, it can be challenging to assess the contextual boundary conditions of their work, especially if they are themselves cognitively embedded in the field of their study. As Whetten (1989: 292) requested, ‘Although it is unfair to expect that theorists should be sensitive to all possible boundary conditions, clearly there is value in conducting some simple mental tests of the generalizability of core propositions’. This suggests some modesty when discussing the wider implications of a study. In mature research areas, such as corporate governance, prior literature often provides important insights as to how and why contextual variations matter. For example, corporate governance structures are shaped by regulations and legal traditions, which are well documented (Estrin & Prevezer, Reference Estrin and Prevezer2011; Nee et al., Reference Nee, Opper and Wong2007). On the other hand, the more novel a theoretical construct, the less prior literature can inform the assessment of its contextual boundaries, and the more authors have to draw on general knowledge and common sense in their reflections.
The assumptions underlying the theory and constructs provide good starting points for this discussion. A useful way to structure this discussion is to go step by step through the research process. In many studies, some theoretical arguments at a high level of abstraction can be generalized beyond the specific context, and authors may be able to provide evidence from archival sources for the existence of a phenomenon. For example, in their study of the contribution of returnee managers in China, Cui, Li, Meyer, and Li (Reference Cui, Li, Meyer and Li2015) first claim that certain insights ‘can be generalized: a) the phenomenon exists, b) the phenomenon contributes to firm internationalization, and c) such contribution is contingent on the structural characteristics of the recipient organization’, and provide their arguments in support for this statement.
However, when it comes to the hypotheses and the operationalization of the theoretical constructs, the scope for generalization is limited. For example, Cui et al. (2015) note that ‘the following insights may to a high degree be context specific (i.e., China-specific): a) the magnitude of the direct effect, and b) the specific characteristics associated with different ownership types, especially with local SOEs’. They then outline possible contextual moderating effects that may affect the magnitude of the effect, and discuss how specific characteristics of Chinese ownership types, especially, local government ownership imply that the moderating effect they found may not be replicated in other contexts. This structure of the discussion section allows to clearly separate the likely generalizable from the likely context-specific, and thus to contribute to the general management literature, without making claims unsupportable by the data.
In the discussion of the generalizability, scholars need to step outside the world of their theoretical model (or, the ivory tower) and reflect where their assumptions apply in the real world. In this process, discussions with other scholars – including reviewers – may also help. For example, good reviewers challenge authors’ unsupported claims of general validity giving specific examples that help establishing boundary conditions.
General theory may be the ambition of many authors. Yet, often it may be more appropriate to start with mid-range theorizing around a specific phenomenon or context of interest, rather than deciding at the beginning whether or not their findings are universal or context-bound. Most theoretical advances start by explaining a limited set of phenomena in a specific context. While many ideas in management research likely have a kernel with general validity, empirical tests in one context can rarely be exactly replicated elsewhere. However, through careful examination of the boundary conditions of a study, generally valid theoretical ideas may emerge.
Therefore, ‘universally valid unless proven otherwise’ is normally not a useful default assumption. The world is and remains diverse with respect to many variables that impact management phenomena. By examining the contextual boundaries of theories and evidence, MEE scholars can advance robust, insightful, and relevant scholarly knowledge.