Perceived Factors Affecting the Integration-Responsiveness Framework: An Analysis from Chinese Multinationals Operating in Australia

This study extends quantitative and conceptual studies that have clarified and assessed the underlying factors influencing multinational corporations (MNCs) international business strategy choices relating to global integration and local responsiveness with the use of cross–level and in-depth interviews. Top management perceptions from nine Chinese MNCs (CMNCs) with operations in Australia are detailed and it is argued a contingency approach tends to prevail within firms with organisational, industrial, and environmental contingencies predominating.


Introduction
Global integration (or unification) is a potential source of competitive advantage for organizations undertaking worldwide activities for it can generate efficiencies from specialization of individual units and facilitate substantial interchange among units (Fan, Zhu, and Nyland, 2012).Empirical studies have shown that a key determinant of efficiency or effectiveness for MNCs is the extent to which their various operating units are differentiated and at the same time integrated, controlled and coordinated (see, Bartlett, 1992;Houten, 1989;Laurent, 1986;Lawrence and Lorsch, 1967a, b;Milliman, Von Glinow and Nathan, 1991;Phatak, 1992;Schuler, Dowling, De Cieri, 1993).Appreciating the need for MNCs to be both internationally integrative and locally responsive, scholars have developed a global integration and local responsiveness framework and identified four basic international business strategies (IBS), namely, international, multi-domestic, global and transnational.However, global integration and local responsiveness are general concepts that mean many things to different observers.Therefore seeking underlying factors and assessing the significance of perceived influences determining global integration and local responsiveness have attracted a large body of quantitative studies (e.g., Birkinshaw and Morrison, 1995;Birkinshaw, Morrison and Hulland, 1995;Luo, 2001Luo, , 2002;;Taggart, 1997).By contrast studies adopting a qualitative research design are few and this article seeks to address this deficiency.Also addressed is the fact that when studying factors affecting global integration and local responsiveness prior studies have centred on MNCs from advanced economies and/or their subsidiaries in emerging markets.This concentration leaves open the question of the relevance of OECD centred findings to MNCs from emerging markets and their subsidiaries in advanced economies.The paper has three parts.It begins with a review of the literature review relating to the underlying determinants identified by previous empirical and conceptual studies.In the next section research questions are developed.Finally implications are drawn from data analysis, and limitations and directions for future studies are advanced.

Empirical Studies of the Integration -Responsiveness Framework
Although it owes its development to Prahalad (1975), Doz (1976Doz ( , 1980)), and Prahalad and Doz (1987), important determinants underlying the global integration-local responsiveness framework (I-R framework) still need to be understood (Bartlett and Ghoshal, 1989); likewise, the underlying determinants of the global integration -local responsiveness paradigm are not necessarily homogeneous (Doz and Prahald, 1991).Responding to their calls, the determinants of the I-R framework have been subject to vigorous empirical evaluation by, among others, Jarillo and Martinez (1990), Roth and Morrison (1990), Johnson (1995), Taggart (1997), andLuo (2001).This particular review of the determinant of the I-R framework in the IB literature follows Ricks, Toyne and Martinez's (1990: 220) objective of identifying "the recent literature on the basic issues so that the reader can be brought up to date and guided towards what can be read in order to obtain the depth of understanding desired." Table 1.Overview of the findings of the empirical studies which focused on the I-R model Note: + = positively related to either global integration (I) or local responsiveness (R); -= negatively related to I and / or R; s= significant; ns = not significant; nt = not tested; ^= non-significant to the other dimension.The sign of "*" means standardization of market demands.
As one focus of the present study is to find the determinants of the global integration and local responsiveness dimensions, empirical studies concerning such determinants will be reviewed.Table 1 contains an overview of the findings of these prior empirical studies (altogether fifteen papers).Selecting the appropriate articles was undertaken by utilizing the two largest academic databases (Business Source Premier, with over 8,000 journals, and Proquest 5000 International, with 5,500 journals) and searching key words in two defined areas: the I-R framework and its determinants or factors.Table 2 displays some details (e.g., authors and publishing time) and the research design (e.g., sample size and their perspectives) of each of these papers.By observing Table 2, several patterns can be identified.First, empirical studies that concentrate on finding the underlying factors of the I-R framework date from the early 1990s.Second, attempts to provide substance to the I-R framework have continually drawn IB scholars' attention.Third, the above fifteen empirical studies provide insights on the determinants of the two salient dimensions.The findings of the empirical studies support the contention that the selection of an IBS should be a function of the contextual requirements, and distinct competitive attributes are found to be emphasized by each group of businesses that prefer a particular strategy.Fourth, the 15 empirical studies mainly focus on integrating the I-R framework with MNCs from advanced economies.In other words, it is difficult to find previous empirical studies of the I-R framework in relation to MNCs from emerging or transition nations.Although regions covered by Luo (2001 and2002) and Yu (2005) are in developing countries, the target population of these studies only involves subsidiaries of MNCs from developed countries.
When conducting empirical studies of the I-R framework, scholars adopt different perspectives, which can be classified into three categories.The first is to view the I-R framework from subsidiary level (Birkinshaw and Morrison, 1995;Jarillo and Martinez, 1990;Luo, 2001 and2002;Taggart 1997;Venaik, Midgley and Devinney, 2000;Yu, 2005).The second perspective is to examine the I-R framework at the headquarter level (Birkinshaw, Morrison and Hulland, 1995;Kim, Park and Prescott, 2003;Johnson, 1995;Roth and Morrison, 1990;).The third perspective does not distinguish between headquarter and subsidiary, rather scholars (Grein, Craig and Takada 2001;Kobrin 1991;Johansson and Yip 1994;Paik and Sohn, 2004) treat MNC's headquarters and all subsidiaries as a whole when observing the determinants of the I-R framework.
The three perspectives identified from Table 2 do not only imply the source of data or fieldwork, but more importantly disclose three feasible approaches for conducting research in relation to the I-R framework.For instance, by conducting survey research at the subsidiary level, Luo (2001) probes determinants of local responsiveness while his later study (2002) investigates underlying factors of global integration.Likewise, through observing MNCs as a whole, Kobrin (1991) focuses on the determinants of global integration, while Grein, Craig and Takada (2001) identify several factors affecting local responsiveness.These empirical results imply both global integration and local responsiveness can be examined from all three perspectives, suggesting researchers have flexibility in terms of data gathering.

Factors Found from Conceptual Studies
Along with empirical studies, there are conceptual studies which are either related to the I-R framework or relevant to one of two dimensions in the IB literature.Table 3 provides an overview of the findings of these conceptual studies.
Table 3. Overview of the findings of conceptual studies or local responsiveness.The authors explain each factor by using the example of Corningware.Hou (2006) also mentions several factors that have influence on the two dimensions.Compared with the earlier findings of Prahalad and Doz (1987), factors identified by Hou (2006) are similar but broader.For example, whether the MNC relies on global clients rather than clients in the host country is one of the factors impacting on global integration (Hou, 2006).
Neither the conceptual studies of Gupta and Govindarajan (1991) nor Chen and Cannice (2006) are directly related to the I-R framework, but the two studies do identify some relevant factors that have impact on the degree of global integration and/or local responsiveness.In the study of Gupta and Govindarajan (1991), the authors propose that if the magnitude and scope of knowledge creation is expected from a subsidiary, the subsidiary should have a higher level of autonomous initiative which further implies a higher level of local responsiveness.Gupta and Govindarajan (1991:785) also conclude that greater decentralization permits managers in subsidiaries "to exercise greater discretion in dealing with the demands of his or her relevant task environment.In contrast to Gupta and Govindarajan (1991), Chen and Cannice (2006) only consider the dimension of global integration.Global sourcing and management control expatriates are the two factors indicated by Chen and Cannice (2006).For example, if a MNC can purchase components or raw material from their parent or peer subsidiary, it is clear that the MNC can potentially increase the level of global integration.Moreover, Chen and Cannice (2006) note that expatriates, especially long-term expatriates, can help HQs to control subsidiaries and implement HQ strategies.

Research Questions Development
The framework has become one of the most influential of the international business strategy models (Taggart, 1997) with many researchers trying to examine structural determinants either of global integration or local responsiveness, or the whole model.Some studies suggest that the assessment of these factors is a prerequisite for MNCs making decisions about resource dispersal, strategic control, goal setting, and power delegation for a focal overseas subsidiary (Doz and Prahalad, 1981;Gupta and Govindarajan, 1991;Yip, 1995).
So far the two most comprehensive studies which solely concentrate on the degree of either local responsiveness or global integration were by Luo (2001) and(2002).In Luo's (2001) study, however, the author proposed an important future research direction.First, since factors influencing integration and responsiveness are "not necessarily the same or are such influences definitely inversely related, future studies should incorporate various factors into an integrated model assessing integration and responsiveness simultaneously" (Luo, 2001: 472).Luo (2001) further argues that, by so doing, it could improve our understanding of which factor only affect either the integration or local responsiveness dimension and what factors are common to both dimensions.Along with responding Luo's (2001) call, if recalling Table 1 and Table 3, factors found from prior studies can be simplified as follows (see Table 4).As mentioned before, the prior empirical and conceptual studies have identified 36 factors that affect the I-R framework based on the perspective of MNCs from advanced economies.Given the paucity of research on influential factors that affect the I-R framework in the contexts of MNCs from emerging economies, this study is one of the first that attempts to formally propose a new research area in international business and management, that is, to seek perspectives of MNCs from emerging country contexts, such as China, on the balance of global integration and local responsiveness.To achieve this goal, a key research question (RQ) is raised for this paper.

RQ:
What factors influence the I-R framework in the context of CMNCs?
Since the identified 36 factors have been classified into three categories (See Table 4), three sub-research questions (SRQs) are developed accordingly.

SRQ1:
What are forces perceived by mangers of CMNCs to integrate their operations globally?
SRQ2: What are forces perceived by managers of CMNCs to localize their operations in host countries?
SRQ3: What are forces perceived synchronously by managers of CMNCs to both integrate and localize their operations?

Research Design
Up to now most studies related to I-R framework have been conducted through direct investigations in western based MNCs and have utilised quantitative rather than qualitative methods.Also most studies have been concerned primarily with the relationships between one or several particular determinants and examined either the degree of global integration or local responsiveness.There remains a dearth of research that explores problems and/or phenomena-centred issues (e.g., MNCs from developing countries investing in advanced economies and the relevant impact on IB), which would reveal cultural distinctiveness of the way MNCs (in this study CMNCs) reconcile global integration and local responsiveness.Moreover, Chinese multinationals operating in advanced economies have not existed long enough to attract a large body of research and the literature presents few studies in relation to the application of the I-R framework by Chinese multinationals.Furthermore, as the research questions, stemming from the extant literature focus on what, how and why of factors affecting the I-R framework in the context of CMNCs, this study is exploratory and non-hypothesis driven, and aims to accurately portray the characteristics and insights of a phenomenon that is as yet to be fully studied (Schmitt and Klimoski, 1991).Therefore, the research methodology applied to this study is qualitative.
Adopting qualitative methodology in this study also has methodological significance in the IB research.Many IB scholars (e.g., Marschan-Piekkari and Welch, 2004;Werner, 2002) have argued it is necessary to reverse the trend of underestimating the value of qualitative research and bring these methods back into the mainstream IB research for several reasons.First, qualitative research has always had a place in the field of IB (Marschan-Piekkari and Welch, 2004).Many important early contributions in the IB field, such as the Uppsala model (Johanson and Vahlne, 1977;Johanson and Wiedersheim-Paul, 1975), and Bartlett andGhoshal's I-R framework (1989 and1998) were built explicitly around qualitative insights (Marschan-Piekkari and Welch, 2004;Birkinshaw, Brannen, and Tung, 2011).Second, as IB lacks the sophisticated theory development of a mature discipline, it requires more exploratory and theory-generating research rather than empirical testing (Marschan-Piekkari and Welch, 2004;Mendenhall, et al., 1993).Third, qualitative research goes beyond the measurement of observable behavior (the 'what'), and seeks to understand the meaning and beliefs underlying action (the 'why' and 'how') (Buckley and Chapman, 1996;Marschan-Piekkari and Welch, 2004).Fourth, qualitative research in IB is particularly preferable when examining other organizations and societies "on their own terms" rather than directly imposing one's culturally bound concepts and theories (Boyacigiller and Adler, 1991: 281), and when random samples may be lacking and respondents may be unfamiliar with questionnaires (Marschan-Piekkari and Welch, 2004).
In addition, this paper adopts multi-case design.Due to the complexity of the internationalization process and/or ownership and industry classification, one particular CMNC as a single case cannot represent the image of the whole.Further, the evidence from multiple cases is more persuasive, and the overall analytical results are commonly regarded as more robust (Herriott and Firestone, 1983).Finally, multiple cases enable the researcher to conduct multiple experiments applying 'replication' logic (Yin, 2003a).A multiple case study design is therefore more appropriate for this study.

Data Collection
In this study Chinese multinationals operating in Australia were chosen for the analysis.China's offshore investment is highly concentrated in a handful of developed countries, namely, the US, Canada and Australia.
The US and Australia are top two recipients of China's outward foreign direct investment (FDI) among twenty-five of the largest recipients globally (MOFTEC, 2003).By the end of 2001, about 30% of Chinese government approved outward FDI went to the three countries, each accounting for 13%, 9% and 8% respectively (Yang, 2005).According to the updated official report (MOFCOM, 2007), almost half of CMNCs' subsidiaries were located in Hong Kong, U.S, Russia, Japan, United Arab Emirates, Vietnam, Australia and Germany by the end of 2006.Through the available data (e.g., from OECD and MOFCOM), the USA received 81% of regional flows to North America, and Australia captured 90% of regional flows to Oceania between 2003and 2006(OECD, 2008).This comparison indicates that both the US and Australia are still the dominant and major recipients of FDI in their regions during the last four years.
The data collection of this study has designed in the way to address our concerns of the impact of industry types, ownership characteristics, entry mode strategies, geographical locations on CMNCs, home and host country perspectives on the I-R framework.As more industries chosen, the more comprehensive view would be obtained on how MNCs reconcile the need to be both globally integrated and locally responsive, to distinguish industry is useful and relevant in choosing the case MNCs.Hence in the study four industries that represent the majority of CMNCs operating in Australia were chosen, namely, (a) energy/mining/resource, (b) financing/banking, (c) trading, and (d) manufacturing.Chinese state owned multinationals have been major players in terms of Chinese outward FDI, because state owned enterprises (SOEs) have a high political profile and rapid resource mobilization, are pivotal in initiating outward investment in the government-controlled investment scenario (Zhang and Van Den Bulcke, 1996).It is clear that case organisations must involve Chinese state owned multinationals.However, there have been more private Chinese companies going abroad in recent years.In Australia, private Chinese multinationals have established businesses in many industries.Therefore, data collection could not simply ignore the existence of private Chinese multinationals.The inclusion of private CMNCs also helps to illustrate how organizational contingencies played a role in the I-R framework based on the difference of ownership.
According to prior studies (e.g., Cui and Jiang, 2009), CMNCs expanding overseas prefer a wholly owned subsidiary (WOS), as against equal and minority joint venture (JV) entry modes.They do so, because full ownership provides greater control over their overseas affiliates (such as high levels of control and integration).However, this view comes mainly from an institutional or strategic behaviours perspective (Cui and Jiang, 2009), rather than considering influence of other theories, such as stage approach and contingency theory and the existence of either entry mode adopted by CMNCs operating in Australia.Moreover, an examination on the entry mode choice of CMNCs in Australia failed to show that either WOS or JV dominated.Hence, this study involved CMNCs that adopted either WOS or JV to enter in Australia.Moreover, selection of case organisations cannot be limited to one or two Australian states.The more states are covered, the more comprehensive view of CMNCs operating in Australia that can be obtained.On the other hand, location advantage is one of three conditions of Dunning's eclectic paradigm (Dunning, 1981(Dunning, , 1993)), and location advantages involve factors such as low transport costs, low input costs, resource availability, infrastructure, and economic and political stability.
Based on the extant literature, these factors have an impact on the I-R framework.Therefore, consideration of geographical location of CMNCs in Australian and selecting case organisations operating in different states would assist in addressing the relevant research questions.
To obtain data the researcher must be "on-site (where the programming is happening) observing, talking with people, and going through program records" (Patton, 1990: 244).From July to November, 2008 nearly five months were spent in the fieldwork conducted in several cities in China (e.g., Beijing, Tianjin, Shanghai, and Taiyuan) and Australia (e.g.Brisbane and Melbourne) where CMNCs headquarters and subsidiaries are located.
Prior empirical studies often collected data from three sources: the firm as a whole, or headquarters only, or subsidiaries only.Although the flexibility exists for examining the I-R framework, it may contain potential limitations.For example, when Luo (2002) examined subsidiary managers' perspectives to observe underlying factors of global integration, the author admitted the findings needed to be further justified by managers from headquarters.Therefore, although exploring the underlying factors of the I-R framework can be done by any of the three sources in the literature, combining observations from both subsidiary and headquarters may provide a more comprehensive and justified view.
In this study, all the interviews of each case CMNC were conducted at both the headquarters in China and subsidiaries in Australia.Over the data collection period, 23 semi-structured interviews were conducted at nine Chinese multinationals (see, the details in Table 5).Of the 23 interviewees, twelve were senior executives at the headquarters level and they held positions ranging from chief executive officer (CEOs), vice-general manager, chief financial officer (CFOs), Chief operation officer (COO), to international project director or executive member of the board of directors.Of the eleven executives interviewed at the subsidiary level in Australia, nine held either the CEO or general manager positions.The other two interviewees were project or departmental directors who were considered senior executives because they played important roles in the strategic decision making process.All participants were interviewed either face-to-face, or through pre-arranged telephone communication interviews.The length of interviews varied from 45 minutes to one and a half hours.Where the participants agreed interviews were recorded by digital recorder.The researcher took notes by pen when interviews could not be recorded.Also notes of interviews, including all the details discussed and any corresponding views expressed by the interviewees and impressions of the researcher were written up without delay.As the majority of interviews were conducted in Chinese, the transcript was recorded in Chinese, and then translated in English.All the data were analysed on the basis of transcripts in English.The researcher subsequently sent transcripts of the interviews to the interviewees for comments, and any feedback was incorporated into the updated transcripts.AS11 Notes: For reasons of confidentiality, the case CMNCs and the interviewees are coded so as to guarantee anonymity of the participants involved.CMNC stands for Chinese multinational corporations.HQ stands for headquarters of Chinese multinationals.AS stands for Australian subsidiaries.

Data Analysis and Findings
Data analysis consists of examining, categorizing, tabulating, testing or otherwise recombining all evidence to address the initial propositions of a study (Yin, 2003a).Moreover, qualitative data analysis has been described as "an array of interpretative techniques which seek to describe, translate and otherwise come to terms with the meaning, not the frequency, of certain more or less naturally occurring phenomena in the social world" (Van Maanen, 1983: 9).Since the preliminary data analysis began with the data collection process, the analytical strategy of relying on theoretical orientation was primarily employed.This strategy guided the inductive procedure of the data analysis and followed the sequence: (a) coding frame; (b) category development; (c) identification of key themes; (d) summarizing findings and developing explanation.We were interested on the interviewees' perceptions on the two dimensions of the I-R framework and the impact of each factor on the two dimensions.We categorized key phrases of each questions in four groups, namely, significant, moderate, not important and not applicable (N/A).For example, some key phases, such as "very important", "a significant issue", "yes, indeed", "very clear", "considerable" are categorized as evidence of rating their perceptions as "significance", while respondents' comments such as, "to some extents", "should be", "in some cases", "play certain roles" are grouped as "moderate".Moreover, at our analysis stage, the difference between "not important" and "N/A" were explicitly distinguished.The following sections on the interpretation of centralization and factor exploitation at both headquarters and subsidiary levels of MNC9 and MNC4 can be served as good examples.

 Centralization
HQ12 from MNC9 at the HQ level commented: As you see, we are a big exporting corporation, and we have 36 major overseas subsidiaries, we couldn't spend our energy on keeping an eye on so called 'top-down' management.If you ask whether it is an issue to global integration, yes, it is.However, in our corporation, we look at each subsidiaries' real performance.So it is not important in terms of considering centralization as an issue for integration.(Lines 158-162; Beijing, China).
AS11 from MNC9 at the Australian subsidiary level believed: Not really, well, it is not important for global integration.For example, I'm not sure about other Chinese companies, but I have certainly had power to run our businesses in Australia.We are operating a 'whole set equipment or project based' exporting business, so performance is important, not those bureaucratic things in this company.If any (business) opportunities relate to Australia, our top management will ask for my opinion (Lines 97-100; Melbourne, Vic., Australia).
Based on the two answers from MNC 9, we deemed 'centralization' 'not important' at either the headquarters perspective or the subsidiary level.

 Factor exploitation
HQ7 from MNC4 at the HQ level explained: No, this is not relevant to our corporation.We never considered if we could find other business opportunities in Australia, such as enter the insurance industry.We can't because it is limited by our business licence.We can only undertake commercial banking.If local governments support us, we can only open branches in other locations, but still in commercial banking and finance (Lines 316-320; Beijing, China).
AS6 from MNC4 at the Australian subsidiary level explained: We don't even think of this one, … neither integration nor responsiveness.Simply, we are not allowed(AS6: Lines 272-273; Sydney, Australia).
Based on the two answers from MNC 4, we ranked 'factor exploitation' as 'N/A' at either the headquarters perspective or the subsidiary level.
Given the above analytical setting, the researcher classified the perceptions from those senior executives at both headquarters and Australian subsidiaries in the three interviewees' perception analysis tables with the colour coding method (Note: each table generated a simplified frequency table during analysis).Notes: Significance rating shown is on the basis of respondents' perceptions.N/A = Not applicable We found the majority of perceived factors affecting the I-R framework in the IB literature are supported by the qualitative analysis results either at HQ/subsidiaries level or at the cross-level.However, perceived factors that are not supported by this study need to be further discussed.First, advertising intensity and premium positioning were not reviewed as significant to global integration at either HQ or Australian subsidiaries level.According to scholars (e.g., Birkinshaw and Morrison, 1995;Korbin, 1991;Venaik, et.al, 2000), these two factors should be deemed significant marketing activities because they can facilitate MNC's global integration in ways, such as unified financial planning, planned and integrated marketing strategies and negotiation with local media.The lack of agreement with previous empirical studies might be due to the business scope of Chinese outward FDI in Australia, and specific products and services provided by Chinese multinationals.For example, although the study involved case organisations (i.e.MNC3) in the manufacturing industry, advertising intensity and premium positioning are not important because their products, such as neoprene rubber and polyvinyl chloride (PVC) in Chinese domestic market and high-density polyethylene (HDPE) and low-density polyethylene (LDPE) in Australian market (MNC3-DOC11), are designed for industrial application rather than for commodity.In the future study, the two factors should be re-investigated by involving a wide range of CMNCs, including commodity manufacturers.
Second, in the theme of global integration factors (e.g., see Table 6) such as country specific advantages (CSAs) and formalization are not supported as significant forces at the HQ level, and not strongly perceived by senior executives from the Australian subsidiaries.CSAs of Chinese multinationals have been discussed by prior studies (e.g., Rugman, 2007;Rugman and Li, 2007), that focus on home country specific advantages and claim that China has relatively strong CSAs such as cheap labour and natural resource, but its multinationals have week firm specific advantages (FSAs) such as lack of brand name, patent and advanced technology.However, the majority of our participants did not believe CSAs help their global integration or even in their internationalization.This finding is possibly explained by the fact that CMNCs are at an early stage of globalization and encouraged by 'going global' policy, and they have not had enough experience to perceive CSAs.For instance, HQ5 (MNC2) explained "in theory our country specific advantages might affect global integration; however I couldn't find a specific example.The influence might be intangible."(Lines: 166-167; Beijing, China).Moreover, it is worth noting that China is still a developing country and its national competitiveness should not be exaggerated in contrast to advanced economies.For example, HQ1 explained: It is hard to say that Chinese firms can benefit from country specific advantages, such as capital, natural resource, labour issues.In some instances, yes, our government's foreign exchange reserve is very high compared with other countries, which in theory can support our going global activities.In fact, although we are big state-owned corporation, it does not mean we have privileges or competitive advantages to easily access the financial support.In terms of natural resource, no, the reason of seeking overseas oil and gas resource is that we [our country] simply don't have enough.Labour issue is complicated unlike in other manufacturing sectors; we can only assign some skilled employees to the project sites [in host countries] (Lines 127-132; Beijing, China).
Within the same company (MNC1), HQ3 argued "according to our experience in some overseas merger & acquisition (M&A) cases, our failure was that we lack home country specific advantages.If we were an American firm, the thing would have already been different....So I would suggest you need to discuss country disadvantage" (Lines 70-75; Shanghai, China).
At the Australian subsidiary level, the majority of interviewees did not support this factor either, but it is worth noting that two interviewees hold different opinions.There is an Australian executive observed -"I think Chinese companies do have CSA.They know where they can reduce cost.They have rich experiences on all kinds of power generation technology; it might be due to their geographic complexity" (AS8,.Although this quote is in line with prior studies, it still cannot confirm the significance of this factor as the cross level design applied in this study.Thus, this has indicated an area for future study. The similar findings happen to the factor of formalization.At the HQ level, the majority of interviewees did not think of formalization as a factor driving their global integration concerns (seven out of twelve treated it as 'not important factor' while one third thought it irrelevant to their organisations).Both state-controlled and private CMNCs did not count the written policies or general guidelines as a method of vertically coordinating subsidiaries.At the Australian subsidiary level fewer interviewees supported this factor, but one of the minority's responses should be noted.
Yes, we followed some rules, procedures, such as daily reporting; and introducing some documented guidelines of operating some special equipment, which are from our headquarters.I think it helps the company's integration internationally (AS9, Lines 115-118).
The above quote shows evidence that the HQs of CMNCs do introduce rules and procedures to their overseas subsidiaries, which is inconsistent with prior perspectives on Chinese firms (e.g., Cai, 1997;Leung, 1995); but the interesting finding is the majority of CMNC executives considered formalization as a "small issue" (HQ3, Line: 135) or "part of routine operations" (AS11, Line: 133) and did not realize that such activities could impact on global integration.On the other hand, it also questions the claim in the extant literature whether it is a "driving force" in the global integration theme.
Third, by observing Table 7 in the local responsiveness theme, the significance of the two factors -business specificity and component localization are not supported at the cross level.Among 12 top managers at the HQs level in nine case organisations, no one asserted that the business specificity had a significant impact on local responsiveness.A typical comment is "I don't think our business or our Australian subsidiary's business is special".However, Australian subsidiaries' managers from seven out of nine case CMNCs think their businesses are unique, which further leads them to think of its impact on local responsiveness.
For example, AS4 commented that "the reason of taking this mining project from its prior owner -a French company is, we found that the quality of the ore is similar with some of our major mines in China, but we still need to learn from this type of ore because they are not exactly the same" (Lines: 100-102; Brisbane, QLD, Australia).He then further explained "because of the specificity of our business, we certainly need to consider some environmental issues, such as Australian natural environment evaluation standard, which is also very different from ours at home, … the impact of our business specificity on local responsiveness is very high" (Lines: 234-239; Brisbane, Australia).
In regard of component localization, the majority of interviewees did not think it is significantly important for their local responsiveness because of their corporate policy -global sourcing, which required them to exploit global efficiencies in the delivery of a product or service.Only one interviewee (out of twenty three in total) thought component localization would be important one then only if it can "build better relationships with local governments and suppliers" (AS1, line 213; Perth, Australia).Additionally, unlike countries such as the U.S., Australian governments do not have strict FDI requirements in terms of regulating component localization.

Checks on Validity
When discussing issues of validity, we need to address both internal and external validity (Tharenou, Donohue and Cooper, 2007;Trochim, 2001).Internal validity concerns the accuracy of the results, while external validity is related to generalizability of the findings.A threat to internal validity arises through researchers' interpretations (Tharenou et al., 2007), and thus we avoided inputting our own values, biases and assumptions into the interpretation.In addition, techniques such as respondent validation and triangulation (i.e., a cross-level comparison, involving different industries and two and more interviewees within one organization, and double check their perception, use more questions related to the same topic to obtain their feedback) were used to assess internal validity in this study.
In terms of external validity, this study sought internal generalizability rather than external generalizability.In other words, the study sought analytic generalization but not statistical generalization (Yin, 2003a).In contrast to external generalizability which refers to generalizability beyond the setting or group, internal generalizability refers to the generalizability of a conclusion within that setting or group studied (Maxwell, 2005).Many methodologists assert that internal generalizability is a key issue for qualitative case studies (Maxwell, 1992(Maxwell, , 2005;;Tharenou et al., 2007).This study adopted several methods in order to enhance the internal generalizability.First, although Lee (2007) claims that a single case study is sufficient to assess the generalizability of findings; Sommer and Sommer (1991) hold a view that increasing generalizability can be achieved by undertaking multiple case studies of the phenomena of interest.Hence, the research design of this study adopted multiple case studies.In doing so, the study investigates whether the development of a theory can be extended to other contexts (Becker, 1991;Yin, 2003a).Second, Tharenou and her colleagues (2007: 82) point out that generalization has to do with extrapolation to theoretical propositions rather than to populations.Therefore, this study emphasizes the existence of particular processes rather than selecting the case MNCs based on different characteristics to show they are typical.Third, according to Hartley (1994), the way that we have explicitly processed --to use the existing literature -can check the generalizability of the findings.

Implications, Future Research and Limitations
By addressing the main RQ and its three SRQs in this paper, perceived factors in our findings can be classified into the three groups, namely, global integration factors, local responsiveness factors and hybrid factors.However, "the same object can often be observed from several different perspectives" (Giere, 2006: 92).In addition to explore which factor is perceived to have impact on the I-R framework in the context of Chinese multinationals, we need to know how these factors have been viewed.In our interviews, phrases and/or words that appeared most frequently are -"relevant to our organisation", "not relevant this industry" and "it depends on …", which indicate these perceived factors are "contingency factors", and also identify which aspects of organisation they effect.This specific qualitative method -the coding analysis -leads us to think of the contingency theory of organisation.As "a major theoretical lens used to view organizations" (Donaldson, 2001:1), through our examination, the 34 perceived factors (excluding advertising intensity and premium positioning) can be treated as multi-contingencies, which can further be classified into three broad aspects, namely, organizational contingencies, industrial contingencies and environmental contingencies (see details in Table 9).For example, factors listed in Tables 6-8, such as centralization, heterogeneity of executive group, and control flexibility are differentiated among organisations; therefore, the impact of these factors on either global integration or local responsiveness depends on organisations characteristics.Likewise, to what extent standardization or competitors' action is important to affect global integration is differentiated between industries.Given nine different industries involved in this study, Tables 6-8 show clear evidence of the industrial difference play significant role in managers' perception on factors affecting the I-R framework.Network/Government Support Environment as a contingency can be dated back to the early research of contingency theory (e.g., Lawrence and Lorsch, 1967a, b;Lorsch and Morse, 1974;Thompson, 1967;Tung, 1979).That is, environmental forces have long been served as contingencies in the related studies.Likewise, organization is also a popular contingent in the literature (e.g., Lorsch and Morse, 1974).Moreover, as Donaldson (2001: 187) identified, "in contingency theory the contingency includes the environment, but is a wider concept that extends to 'context' more broadly, thereby encompassing intra-organizational variables…" By observing the first column in Table 9, over half (19 out of 34) factors identified from previous studies have organizational characteristics.Hence, these factors need to be treated as organizational contingencies.
In contrast to environment and organization as contingencies, treating industrial factors as contingencies are not common in the literature.However, industry factors have deep root in the literature of IB.Dunning (1981) holds a view that industrial structural imperfections in foreign markets not only make FDI preferable to trade or licensing but also determine the relative attractiveness of some host countries over others or the home country itself.An industry's structural forces determine the conduct, behaviour, and strategy of firms in that industry (Scherer and Ross, 1990).Additionally, Luo (2001) asserts that industrial factors have equal importance with organizational and environmental factors in terms of studying the I-R framework.Furthermore, contingency theory holds that if the fit can be reached by contingencies, then higher performance results (Donaldson, 2001).Arguably, without the benchmark of industry, it would be difficult to assess to what extent the higher performance results.Hence, factors with industrial characteristics have been separately categorized as contingencies.
Moreover, the contingency approach means that the effect of one variable "X" on another "Y" depends upon the third variable "Z" (Arnold, 1982;Donaldson, 2001).In relation to the I-R framework, "X" refers to the two dimensions (the degree of global integration and local responsiveness) while "Y" means the four international business strategic options.Clearly, the "Z" involves the three groups of contingencies.In the further studies, especially when quantitative research method is followed as another research method, it should note there are two levels of assessment in regard of the factors.If we disregard the strategic implication of the I-R framework (that is, remove the four IBS), factors in Tables 6-8 demonstrate their impact on the degree of either global integration or local responsiveness.Therefore, the assessment should focus on the extent to which some or all of them have impact on the degree of global integration or local responsiveness.However, if the IBS are involved, factors can be deemed as contingencies (see Table 9).Thereby, for identifying which IBS is more relevant to an MNC, the primary task is to seek "fit" among the three groups of contingencies.
This study extends quantitative and conceptual studies that have clarified and assessed the underlying factors influencing MNC choices relating to global integration and local responsiveness with the use of cross-level and in-depth interviews.The findings in this study contribute to our understanding in the IB field, such as incorporating various perceived factors into an integrated model assessing integration and responsiveness simultaneously, and obtaining empirical evidence of applying the I-R framework in the context of MNCs from emerging market.The conceptual contribution of this study is that these perceived factors can be classified as organisational, industrial, environmental contingencies, which indicates that a contingency theory could be used for future studies in this field, such as seeking strategic fit within the I-R framework.Nevertheless, some limitations of the study need to be addressed.One of the major limitations is that the study is focused on managers' perceptions on factors affecting the I-R framework.Future studies could include multiple sources of data, such as documents and observations.By supplementing information from documents or observations other than only in-depth interviews can further enhance the validity of data analysis.A cross-level research design in this study only involves Chinese multinationals' HQs and their Australian subsidiaries.In the future studies, a comparison between HQs and their subsidiaries located in two or more countries and even with subsidiaries from home country could be considered (Fan et al., 2012).In doing so, the underlying factors, such as culture distance and domestic competition can be better explored.A cross-level design is therefore warranted.In addition, the majority case organisations involved in this study are big Chinese multinationals that are dominant at least in their industries.The future studies could involve more multinationals that are small and medium enterprises (SMEs), which in turn could improve our understanding on organisational contingencies, and how different managers' perceptions can be between large multinationals and SMEs in regard of environmental contingencies, and their foreignness in overseas markets.

Table 2 .
A summarised table of empirical studies associated with table1

Table 4 .
Categorizing factors found from prior studies

Factors only affecting global integration Factors only affecting local responsiveness Hybrid Factors
this category can either affect the degree of global integration or the degree of local responsiveness, or synchronously drive the both dimensions.To date, little research has mentioned the hybrid factors in study of the I-R framework, but this article has demonstrated the existence of hybrid factor based on literature review.It is important finding because it can potentially deepen the understanding of these underlying forces in relation to the I-R framework.

Table 5 .
Case CMNCs and the interviewees

Table 6 .
Perceptions on factors affecting global integration

Table 7 .
Perceptions on factors affecting local responsiveness Significance rating shown is on the basis of respondents' perceptions.N/A = Not applicable

Table 8 .
Perceptions on hybrid factors

Table 9 .
Re-categorizing factors based on contingency perspective