Effects of nationality on global strategy.

Does corporate nationality affect the use of global strategy? Since most definitions of global strategy (e.g., Porter 1986, Bartlett/Ghoshal 1989) involve integration of policies, operations, and personnel across nations, the nationality of the company concerned should a priori have some effect. But nationality is a multidimensional phenomenon (including citizenship, history, culture and experience) and can apply to different aspects of an MNC (including the past and current location of corporate headquarters, the nationality of managers, and the national location of units and subsidiaries). The formulation of global strategy lies primarily in the domain of business or corporate headquarters, and even its implementation is usually driven by headquarters. Indeed, overseas subsidiaries tend to resist global strategy because of its attendant loss of autonomy. So in this study, we focus on the effects of the nationality of the parent company. Parent nationality can differ from the nationality of the business unit (e.g., French Thomson now owns American RCA) and from that of senior managers (e.g., a Cuban as the CEO of Coca-Cola or a German, and then a Frenchman, as the CEO of Apple Computer), but these are exceptions, albeit increasing, and a fruitful topic for another study. Another exception is that the company may be registered in one country, but have its senior management located in another. For example, Jardine Matheson moved its registration to Bermuda, but most of its head office personnel continue to operate out of Hong Kong. For the vast majority of MNCs, the original and current countries of corporate headquarters and of most business headquarters are still the same. So we define nationality as the country in which most HQ managers of the parent company are located.

Parent nationality can affect global strategy in a number of ways. First, national history has affected the general order of entry into international markets: Western European first, Americans second, Japanese third, and other Asian countries and some Latin American ones going international most recently. Early entry clearly leads to greater global dominance, it being well established that order of entry is one of the primary determinants of share. The timing of entry should also affect global strategy. In particular, the standardization and specialization in global strategy requires lower trade barriers and differences in national customer tastes. More recent international entrants, like Japanese companies in the post-WWII period, have faced increasingly lower trade and taste barriers.

Second, parent nationality affects what types of organization structures, management processes, personnel practices and managerial culture predominate. Some of these differences may arise from the accidents of development and others may be rooted in national culture or be a combination of the two.

Existing Evidence on Role of Nationality

Many studies in international business have found nationality effects and some have found no effects. But in most cases, these studies have not been concerned with global strategy, as we define it in terms of the extent of uniformity and integration across countires. These earlier studies have examined within-country strategies and organizational approaches. For example, Douglas and Craig (1983) found no important differences, using PIMS data, in explaining the performance of US and European national businesses within individual countries. The few studies looking at global strategy have found national differences, and we cite them in the next few paragraphs.

Some overall characterizations of possible national differences in globalization can be made, particularly for the three regions that have dominated global business - Western Europe, the Unites States and Japan. In their response to globalization forces, MNCs in these three regions are likely to be affected by their history of internationalization, to some extent following Stopford and Wells' (1972) stages of international development.

European MNCs

Following their colonial flags, or imitating neighbors in the case of countries without colonies, European MNCs were often the first to internationalize, many beginning in the 19th century. In this early period of poor transportation and communications, and high trade barriers, European MNCs typically built up overseas subsidiaries that enjoyed a large degree of autonomy. Many European MNCs are now finding that this local autonomy greatly inhibits the ability to formulate and implement globally integrated strategies. For example, Philips N. V.'s problems with its overly-independent subsidiaries have been well documented (Aguilar/ Yeshine 1988, Jeelof 1989). Both perceptions and the reality of their many historical, cultural and political differences may spur Europeans to favor local rather than global solutions. A study over the time period, 1972-1987, found that national or local businesses performed better than continental ones in Europe, while the reverse held in North America (Yip 1991). A study of the European food industry in the 1980s found that European companies mostly marketed national brands while it was left to American companies to succeed with pan-European ones (Bennigson 1992). On the other hand, because of the small size of their home markets, European companies tend to have a very high percentage of revenues being international, and also have a long history of moving managers around the world. The overall globalization challenge for the typical European MNC seems to be to move from the direction of local autonomy toward global integration.

American MNCs

Perhaps the key distinguishing feature of American MNCs is their huge home market. This large domestic base has frequently resulted in the building of separate US and international divisions, with a large wall between them. Another consequence has been a tendency to under-adapt for local markets, utilizing American products and programs rather than tailoring for international markets. This orientation toward the home market also makes it difficult to achieve a global strategy that best exploits opportunities around the world and leverages worldwide capabilities.

Japanese MNCs

Later than Europeans and Americans to internationalize, most Japanese MNCs have until recently pursued an export-based model, with its concomitant centralization. While some studies have found a high degree of local autonomy for Japanese subsidiaries, Putti, Kulwant, and Stoever (1993) concluded that this autonomy is apparent rather than real. The extensive use as subsidiary managers of Japanese nationals, all of whom have been long socialized in the corporation back home, and the consensual style of Japanese management mean that Japanese executives act not so much as managers of quasi-independent subsidiaries but as the parent companies' representatives. Some aspects of this centralization favor global strategy, particularly in terms of integration. Other aspects hurt global strategy, as in the lack of a globally specialized network, like that advocated by Bartlett and Ghoshal (1989). As cost pressures and trade frictions spur Japanese MNCs to relocate manufacturing activities, they face the challenge of preserving the advantages of their unique system.

These historical differences are likely to affect all aspects of globalization, from perceptions of industry globalization drivers to the actual use of global strategy. Various studies have found differences in organization structure and management processes among American, European and Japanese companies (e.g., Franko 1971, Jaeger 1983, Egelhoff 1984, Abegglen/Stalk 1985, Tsurumi 1986), although mostly without relating these differences to global strategy. Bartlett and Ghoshal (1989) argue that American MNCs stress formalization of structure and processes to achieve global integration, European MNCs stress socialization, and Japanese MNCs centralization. This study attempts to directly estimate the effect on global strategy of such national differences in organization and management.

Determinants of Global Strategy

The role of nationality in global strategy can best be examined in a complete model that includes other determinants of global strategy. We adapt and extend a model developed by Johansson and Yip (1994). Our adapted model of global strategy has the following constructs:

* Business global strategy

* Industry globalization drivers

* Organization and management

- Global organization structure

- Global management processes

- Global human resources 

- Global culture

* Business and company characteristics

- Business global market share

- Business extent of internationalization

- Size of corporate parent

Business Global Strategy

We define global strategy as strategic choices (strategy content) that have the characteristic of being globally uniform or integrated (Yip 1992). We view five aspects of global strategy to fully operationalize this construct:

* market participation strategy

* product strategy

* location strategy for the value chain

* marketing strategy

* competitive moves

While other aspects might be added, they would mostly overlap with the above list of five. In addition, we treat all organizational and managerial aspects of global strategy as separate constructs, being enablers of strategy rather than choices of strategy content.

Industry Globalization Drivers

Following industrial organization theory, we propose that strategy choices and outcomes are significantly determined by industry conditions. In the case of global strategy, these conditions are industry globalization drivers.

The drivers identified in the literature can be grouped into four categories, each based on a different set of actors: market (customer base), cost (firm base), government (regulator base) and competitive (competitor base). Together they can operationalize the construct of industry globalization drivers, as follows:

* Market globalization drivers comprise market- and customer-based conditions that favor the use of global strategy, and are indicated by the global commonality of customer needs and tastes, the importance of global customers and channels, the global transferability of marketing approaches (reaction of local customers to foreign marketing, and the effects on customers of exposure to product and marketing experiences while overseas).

* Cost globalization drivers comprise cost-based conditions that favor the use of global strategy, and are indicated by global scale economies, steep experience curve effects, sourcing efficiencies, favorable logistics, differences in country costs (including exchange rates), and high product development costs.

* Government globalization drivers comprise government-imposed conditions that favor the use of global strategy, and are indicated by tariff and non-tariff barriers, compatible technical standards, and common marketing regulations.

* Competitive globalization drivers comprise competitive conditions that require the use of global strategy, and are indicated by the extent to which competitors use global strategy.

Organization Factors

Organization factors such as structure, management processes, people and culture clearly affect how well a desired global strategy can be implemented, probably more so than for most other types of strategy (Prahalad/Doz 1987, Ghoshal 1987, Bartlett/Ghoshal 1989). Roth, Schweiger and Morrison (1991) found, in a study of 82 business units competing in global industries, that business unit effectiveness was a function of the fit between the international strategy and the organizational design. Johansson and Yip (1994) found, for 36 American and Japanese MNCs, that organization and management factors play a major intervening or moderating role between industry globalization drivers and the implementation of global strategy. There has, of course, been extensive work on the subject of MNC control and coordination mechanisms (reviewed by Martinez and Jarillo 1989), but this mostly predates the current debate on global strategy.

We define global organization structure as formal organizational devices that globally centralize management authority. This construct is indicated by having one manager whose primary job it is to be head of the worldwide business, by whether there is an international division that does not contain the domestic business (if yes, authority is dispersed), and by whether there are managers with global line authority for each major function (research, development, purchasing, raw material processing, intermediate production, final production, marketing, selling, distribution and service).

We define global management processes as processes, both formal and informal, that enable cross-country coordination and integration. This construct is indicated by the extent to which a company has processes for coordinating strategy across countries, the extent to which the business has global budgets that are used for global programs (as opposed to national budgets for national programs), the extent and frequency of meetings involving senior managers from around the world, the extent to which senior managers are evaluated and compensated on the basis of global and not just regional or national performance, and the extent to which a business collects strategic information, such as market share and competitor data, from around the world in a consistent format on a regular basis.

We define global human resources as the use of managers outside their home countries. This construct is indicated by the extent to which foreign nationals hold senior management positions in the home country, home country nationals hold senior management positions in other countries, and foreign nationals hold senior management positions in other countries than the home country.

Various aspects of corporate culture can affect an MNC's ability to develop and implement global strategy. Beginning with Perlmutter (1969), the degree of international or global orientation in corporate culture has been viewed as an important influence on MNC strategy and performance - with Perlmutter distinguishing among ethnocentric, polycentric and geocentric cultures. Another relevant aspect of corporate culture is the degree to which the culture of the home country company is transmitted to international subsidiaries. But dominance by the home-country culture probably does not foster the neutral viewpoint needed in global strategy. Thus, Ohmae (1985) argued that managers and units should be treated as if they were "equidistant from the corporate center". So a culture of a global rather than a national identity should help the implementation of global strategy. We define global culture as a culture that is neutral between countries or nationalities. This construct is inducated by the extent to which the overall culture of the company is global rather than that of the home country.

Business and Company Characteristics

In addition to organizational constructs, some business and company characteristics, not much affected by industry globalization drivers, should also affect both the intent to use and the ability to implement global strategy. These characteristics include a business's global market share, a business's extent of internationalization, and the size of the corporate parent.

Overall Model of Global Strategy

We show the complete system in Figure 1. Our model focuses on explaining Business Global Strategy. Hence we examine all possible paths to that construct while limiting the number of paths to other contructs. Furthermore, the Johansson and Yip (1994) study, of which this is an extension found that paths between the various organization constructs were not significant. In our model, Industry, Globalization Drivers has a direct path to Business Global Strategy, but also indirect paths via Global Organization Structure, Global Management Processes, Global Human Resources, and Global Culture. These organizational constructs also have their own direct paths to Business Global Strategy. The three business and company constructs - Business Market Share, Business % International, and Corporate Parent Size - each have direct paths to Business Global Strategy.

In our model we both adjust for the effect of nationality and test for the strength of this effect by adding a path from a nationality construct to each of the other constructs (paths not shown in Figure 1).

Study Design

Being able to collect good data continues to pose a major barrier for empirical work in global strategy. Problems include those of access and cooperation, interpretation and crossnational comparability. This study faced the additional challenge of requiring data on a very large number of multiple measures for several constructs, as well as perceptions and interpretation from senior managers. Accordingly, we chose personal interviews as the best way to get the attention and length of time needed, and to minimize problems of interpretation.

Choice of Nationalities to Study

With some increasing exceptions, most major MNCs are headquartered in the United States, Japan and Western Europe, matching Ohmae's (1985) "Triad". We treat the United States, Western Europe and Japan as three "national" groups. While some argue that European countries are too heterogeneous to group together (e.g., Hedlund/Nonaka 1993), increasing European convergence and integration suggest that commonalities may be greater than differences. Furthermore, our objective is to discover if there are any uniquely European nationality effects. We test the estimated Europan linkages for heterogeneity using a split sample technique (see below).

Operationalization of Constructs and indicators

To operationalize the constructs and their indicators, we developed a questionnaire that converted each indicator into an item with possible responses on a 5-point scale in most cases. The Appendix describes the questionnaire items used in our final model.

Sample

We recruited a sample of 63 worldwide businesses (defined as business units that marketed in countries on more than two continents, and that had at least 20% of revenues outside the home country) in very large MNCs (19 American, 24 European, and 20 Japanese).

Although the limited sample prevents wide generalizability, this latter was enhanced by targeting the archetypal MNC - very large companies with a high proportion of international revenues. In addition, 63% of the American businesses were the global market leader in their industry, 42% for the Europeans and 30% for the Japanese. Further, 56% of the American businesses, 96% of the European businesses, and 67% of the Japanese businesses were the domestic market leaders. The American businesses averaged 42% of revenues as being international, the Europeans 81%, and the Japanese 36%. The American parent companies had [TABULAR DATA FOR TABLE 1 OMITTED] mean revenues of $19.5 billion, the European $16.5 billion, and the Japanese $11.5 billion.

Data Collection

We used semi-structured personal interviews to administer the questions. In each of the American companies, two managers were interviewed independently (and their responses averaged), and a researcher was present as the questionnaire was completed. In Europe one respondent was interviewed in each company. In Japan, the questionnaires were left with the designated respondent, who then polled the relevant people in the organization across the required functions and reported their consensual responses. This procedure also allowed us to gather extensive qualitative responses, which we used to help interpret our findings.

Analytical Methods

We estimated the model shown in Figure 1 using latent variable causal modeling via the Partial Least Squares technique (PLS). This approach is now quite well established as a method for estimating path coefficients in latent variable causal models (see Fornell/Bookstein 1982, and Fornell/Lorange/Roos 1990). PLS has the advantage over LISREL of requiring less stringent assumptions about the randomness of the sample and the normality of the distribution of variables (Wold 1982). Furthermore, it can accept smaller sample sizes, as each causal sub-system sequence of paths is estimated separately (Anderson/Gerbing 1988). For example, Lohmoller (1982) demonstrates examples where a model with 27 variables is appropriately estimated with only 10 observations and a model of 96 indicators and 26 constructs is estimated on a sample of 100.

In the PLS technique, each construct is represented by a set of latent variables or indicators, these being the ones listed in the Appendix. For the nationality constructs we used a yes/no dummy variable for each nationality. Unlike linear regression, PLS does not provide a unique solution for the third dummy when two dummy variables are used. The solution is to estimate the model three times with each possible pair of dummies, and calculate an average coefficient for each nationality dummy.

An important issue is whether the businesses from the various European countries comprise a sample that is too heterogeneous to jointly represent "European nationality". We tested for this homogeneity by splitting the European set of businesses into two halves, and testing the differences in path coefficients between the two alternate runs (the Japanese and American data staying the same to preserve degrees of freedom). If the European sample were heterogenous, we would have different results from dropping half the observations. This was not the case - the estimated coefficients were very similar between the two runs.

On the suggestion of an anonymous reviewer, we carried out a few more split sample runs to test for homogeneity. Although random splits constitute the traditional procedure for testing homogeneity, one can argue that specific countries or regions will differ from others. To test such a proposition, we made four additional split runs. In one, the six British companies were held out, in another the seven Swedish countries. A third run eliminated the British, Swedish, and Norwegian companies. A fourth eliminated the continental companies (from France, the Netherlands, Germany, and Switzerland). Comparing these runs basically serves to test whether the Scandinavian and the British companies are different from their counterparts on the European continent. In all runs the American and the Japanese companies were kept in so as to preserve sufficient degrees of freedom. The model tested was the one depicted in Figure 1 with Japan and European dummies linked only to Global Strategy in order to preserve degrees of freedom.

The results showed a great degree of homogeneity. The estimated path coefficients in the models varied only slightly between the runs. As could be expected there were some sign shifts for the very small coefficients. For example, the link between Organizational Structure and the use of Global Structure varied between 0.03 and -0.08. But the magnitude of the larger coefficients remained robust, with, for example, the Globalization Driver link to Global Strategy consistently low, and with only slight variations when the Northern or the Continental companies were omitted.

It has to be admitted that these tests for homogeneity of the European companies are not very powerful. Because the majority of the sample (the Japanese and the American companies) stays constant between the split runs, there is a good deal of inertia in the coefficients, and any heterogeneity has to be strong to surface. Given these limited tests, however, the results do support a homogeneity assumption.

Results

In presenting the results, we discuss first the measurement model - how the indicators loaded on to the constructs, before discussing the paths in the model and the nationality effects.

The Measurement Model

We initially estimated the model using a full set of indicators for each latent construct. Then for the sake of parsimony and following standard practice (see, e.g., Fornell 1987), we re-estimated the model for a more limited set of significant loadings. For theoretical reasons, we used a relatively low cutoff level (0.20) for the loadings, statistically significant at the 0.01 level, but much lower than the typical 0.40 to 0.60 range often used in purely empirical applications. The low cutoff enabled us to track the influence of indicators that a priori theory suggests should enter, and gives a sense of the idiosyncrasies of this particular sample of firms.

Table 2. Indicator Loadings

Industry Globalization Drivers

Overall Market Drivers                         0.80
Change in Market Drivers                       0.55
Overall Cost Drivers                           0.29
Change in Competitive Drivers                  0.31
Reaction to Foreign Packaging                  0.27
Overseas Exposure to Promotion                 0.66
Overseas Exposure to Brands                    0.51
Global Scale Economies                        -0.44
Transportation Costs                          -0.71

Global Organization Structure

One Global Head                                0.52
International Division                        -0.79
Average Line Heads                             0.76

Management Processes

Cross-Country Coordination                     0.86

Global Budgeting                               0.58
Global Group Meetings                          0.76
Global Performance Review                      0.67

Global Human Resources

Foreign Nationals in Home Country              0.27
Home Country Nationals in Other Countries      0.49
Foreign Nationals in Other Countries          -0.65

Global Culture                                 1.00

Business & Company Characteristics
Business Global Market Share                   1.00
Business % International                       1.00
Parent Company Size                            1.00

Global Strategy

Actual Overall Global Strategy                 0.90
Standardized Products                          0.39
Uniform Marketing                              0.21
Integrated Competitive Moves                   0.80

Overall Measurement Explained Covariance = 65%

Note: Loadings of 1.00 apply when a construct has only one
indicator. (9/12/95 and 9/21/95)

The loadings for the retained indicators are given in Table 2. The measurement explained covariance of 65% is quite high, showing that the indicators selected provide satisfactory representation of the constructs. Industry Globalization Drivers are dominated by market factors, including promotional spillover effects, and transportation costs. For this particular sample of firms, cost savings seem to be a relatively moot issue. Organizational Structure reflects clearly the apparent need for these global companies to avoid setting up an international division and instead rely on global line managers - but the later relationship to Global Strategy implementation remains weak as will be seen below.

The Management Processes construct, an important determinant of Global Strategy as we will soon see, reflects the effectiveness of cross-country and group-based coordination. As far Global Human Resources, in the end not a very important factor for Global Strategy, the major factor at work seems to be the degree to which foreign nationals are used in other countries. The Global Strategy construct itself reflects mainly the self-assessed globality and the degree of which the company makes integrated competitive moves. Uniform marketing is for these companies a surprisingly weak component of global strategy.

Path Coefficients Between Latent Constructs

Figure 2 shows the path coefficients between constructs in the estimated model, and the [R.sup.2] for each dependent construct. As our objective is to examine nationality effects in a model that has a good explanation of global strategy, we are satisfied with the [R.sup.2] of 0.71 for Business Global Strategy, and not concerned with the lower [R.sup.2] levels for intermediate constructs. In terms of goodness-of-fit, while a Bentler-Bonett value of 0.38 is relatively low in absolute terms, an explained covariation of 55% is acceptable (Bentler 1989, p. 93).

The results show that Industry Globalization Drivers and Global Management Processes are the dominant determinants of Global Strategy, with significant path coefficients (0.38 and 0.41) compared with much smaller paths for the other constructs. Business % International and Corporate Parent Size have smaller, but significant paths, while Global Organization Structure, Global Human Resources, Global Culture, and Business Global market Share have small, insignificant paths. Global Management Processes has not only a significant direct path to Business Global Strategy, but also a significant indirect, moderating effect on the path of Industry Globalization Drivers (0.61x0.41 = 0.25 indirect effect for Industry Globalization Drivers via Global Management Processes).

Nationality Paths

Figure 2 shows the nationality path coefficients (from the "E", "J", and "A" boxes to each construct). Each coefficient is to be interpreted as the "average shift" in the construct due to the business' country of domicile, relative to the other countries. For example, the coefficient for the "Japan effect" on the use of Business Global Strategy is 0.18, to be compared to the two negative values for the US (-0.14) and Europe (-0.03), indicating the greater use of global strategies by the Japanese.

European Nationality has significant, positive paths to Industry Globalization Drivers, Global Organization Structure, Global Culture, Business Global Market Share, Business % International; but no significant paths to Global Management Processes, Global Human Resources, Corporate Parent Size, and Business Global Strategy. Japanese nationality has significant, positive paths to Business Global Strategy; significant, negative paths to Global Organization Structure, Business Global Market Share and Corporate Parent Size; but no significant paths to Industry Globalization Drivers, Global Management Processes, Global Human Resources, Global Culture, and Business % International. American Nationality has significant, negative paths to all constructs except Business Global Market Share, and a significant, positive path to Corporate Parent Size.

Discussion

The primary objective of this study was to examine nationality effects in global strategy, and we defined nationality as the country in which most HQ managers of the parent company are located. Nationality does have significant effects. Overall, the American businesses make significantly less use of global strategy (path of -0.14), the Japanese make significantly more use (path of 0.18) and the Europeans slightly less use (path of -0.03). This end result is consistent with the American businesses having mostly negative nationality effects for the explanatory constructs. The results for the Japanese are as expected, and reinforce our argument that they have benefitted from lateness to bypass the multidomestic stage of internationalization and jump to the global stage. Lastly, the Europeans seem to have adjusted from their first-mover disadvantages of being overly multidomestic. A possible explanation is that their very high dependence on international revenues has made them realize their vulnerability if lagging in globally integrating their strategies. Johansson and Yip (1994) found that American business lagged Japanese businesses in the use of global strategy. This study finds that they lag European businesses also.

For industry globalization drivers, the European businesses face, or perceive that they face, significantly stronger drivers (path of 0.15), and the Americans significantly weaker drivers (path of -0.13), while the Japanese effect is negligible (path of -0.01). We had sought to match the three samples by industry type (see Table 1 earlier) and on an indicator by indicator basis, there are few significant differences by nationality. But it seems that in the context of a complete model of global strategy, nationality differences appear even for industry globalization drivers. The large American home market seems to have reduced the perception of globalization forces and the consequent drive to global strategy.

Although the path from global organization structure to business global strategy is not significant, the nationality effects on this construct are interesting in themselves, supporting prior expectations. The results show that the European businesses have organization structures that are more globally centralized (path of 0.53) than the Japanese (path of -0.40) and the Americans (path of -0.11). The explanation lies in the much lower European incidence (29%) of a separate international division than the Japanese (95%) and the Americans (74%). Similarly, the Europeans have a higher incidence of global line heads for individual functions.

The findings for global management processes (European path of 0.03, Japanese path of 0.08, and American path of -0.10) supports the argument that the Japanese offset their apparently non-global formal organization structure with informal global management processes. Indeed, the effect would probably have been even stronger had we measured the various informal methods of coordination that Japanese are known to use. But as most researchers have found, these informal processes are very hard to discern with formal research methods. It is notable that the American businesses have a significant negative nationality effect for this construct, global management processes, that has the greatest overall influence on global strategy. Again, the large American market may contribute to this relative lack of global management processes.

Although the paths from global human resources and global culture to business global strategy are not significant, the nationality effects on these constructs are interesting in themselves, supporting prior expectations, as with the effects for global organization structure. These results again suggest that the American businesses have the least effective process for global integration. In contrast, the Europeans have both human resource practices and cultures that are most suited for global strategy.

The nationality effects for the business and corporate constructs can be particularly interpreted as control variables. The Europeans have somewhat higher market shares and this is reflected in the significant, positive path from European nationality to this construct. Similarly, the paths for corporate parent size and for the extent of internationalization all match the relative sizes and proportions in the sample.

Europeans love to proclaim their differences from each other and the uniqueness of their particular countries. But the large number of significant European nationality effects suggests a high degree of homogeneity for this group of businesses from large European companies and for the constructs studied. This homogeneity is higher even than within the group of Japanese businesses. So this study finds a greater degree of intra-European commonality than have earlier ones.

Porter (1990) has stressed the vital role of the home country in competitive advantage, while Ohmae (1990) has argued that global firms should be stateless. Our study provides more support for Porters's view: nationality does matter in current approaches to global strategy, although in the case of Europe, it seems to apply at the regional level rather than that of individual countries.

Limitations of This Study and Suggestions for Further Research

Generalizations from our results need to recognize certain limitations of our study. First, our sample is of moderate size and restricted to very large companies. Second, the model used is not completely specified, concentrating only on the part directly related to business global strategy, and it is possible that a more holistic model, requiring a much larger sample, would have changed some of our findings (Bagozzi/Phillips 1982). So a larger sample allowing a holistic estimation technique, such as LISREL, might reveal different relationships. Third, our model cannot prove causation, merely infer it from the estimated relationships (the paths). The use of multiple time periods would strengthen the ability to draw causal conclusions.

Notes

1 This research was sponsored and supported in part by the Marketing Science Institute. The authors thank MSI and the companies that participated in the study, the Global Information Services division of Nielsen Marketing Research and the British Institute of Management for their help with recruiting companies in Europe, and also Professor Ikujiro Nonaka, Institute of Business Research, Hitotsubashi University, and Professor Koji Tsubaki, School of Commerce, Waseda University, for their help with recruiting and interviewing companies in Japan. The Center for International Business Education and Research at UCLA, and NFR/NORAS, Norway (grant No. 2151309), also provided support for some of the research in Europe.

Appendix. Measures of Globalization Used in Model

Industry Globalization Drivers

Overall Market Drivers: Overall strength of market factors favoring globalization in this industry today (1 to 5 scale).

Change in Overall Market Drivers: Direction and speed of change in market factors favoring globalization in this industry category today (1 to 5 scale).

Overall Cost Drivers: Overall strength of cost factors favoring globalization in this industry today (1 to 5 scale).

Change in Overall Competitive Drivers: Direction and speed of change in competitive factors favoring globalization in this product category today (1 to 5 scale).

Reaction to Foreign Packaging: Reaction of customers in different countries to non-local packaging (1 to 5 scale).

Overseas Exposure to Promotion: Effect on customers' loyalty in their home countries from exposure to sales promotions in foreign countries (1 to 5 scale).

Overseas Exposure to Brands: Effect on customers' loyalty in their home countries from exposure to brands in foreign countries (1 to 5 scale).

Global Scale Economies: Global market share needed to achieve economic scale in the total business (% scale).

Transportation Costs: Average cost of transporting products from the United States to Europe (Chicago to Frankfurt) or from Japan to the United States (Tokyo to Chicago) as a percent of sales value (excluding tariffs and duties) (% scale).

Global Organization Structure

One Global Head: Yes/no measure on whether there is one person whose primary job it is to be head of the worldwide business (0-1 scale).

International Division: Yes/no measure on whether there is an international division that does not contain the domestic business (0-1 scale).

Average Line Heads: Average of yes/no measure for whether there is a global line head for Research, Development, Purchasing, Raw Material Processing, Intermediate Production, Final Production, Marketing, Selling, Distribution, and Service (0-1 scale).

Global Management Processes

Cross-Country Coordination: Extent to which the business has processes for coordinating strategy across countries (1 to 5 scale).

Global Budgeting: Extent to which the business has global budgets that are used for global programs, as opposed to national budgets for national programs (1 to 5 scale).

Global Group Meetings: Extent and frequency of meetings involving senior managers from around the world (1 to 5 scale).

Global Performance Review: Extent to which senior managers are evaluated and compensated on the basis of global and not just regional or national performance (1 to 5 scale).

Global Human Resources

Foreign Nationals in Home Country: Extent to which foreign nationals hold senior management positions in the home country (1 to 5 scale).

Home Country Nationals in Other Countries: Extent to which home country nationals hold senior management positions in other countries (1 to 5 scale).

Foreign Nationals in Other Countries: Extent to which foreign nationals hold senior management positions in other countries than the home country (1 to 5 scale).

Global Culture

Global Culture: Extent to which the overall culture of the company is global rather than that of the home country (1 to 5 scale).

Business and Company Characteristics

Business Global Market Share: Extent to which the business was one of the largest competitors, medium sized competitors or smaller competitors (1 to 3 scale).

Business % International: Percent of revenues of the business outside its home country (% scale). Corporate Parent Size: Worldwide revenues of the business's parent company (absolute value).

Global Strategy

Actual Overall Global Strategy: How global the overall strategy is today (1 to 5 scale).

Standardized Products: Extent to which the business's products are globally standardized, i.e., as indicated by percent of cost of product or service that is in components that are standardized (1 to 5 scale).

Uniform Marketing: Extent to which the business takes a different approach in each country or a uniform approach for each marketing element. Average of measures for Uniform Brands, Uniform Packaging, Uniform Absolute Pricing, Uniform Relative Pricing, Uniform Advertising, Uniform Promotion, Uniform Selling Approach, Uniform Distribution Approach (1 to 5 scale).

Integrated Competitive Moves: Extent to which the business coordinates competitive moves across countries (1 to 5 scale).

References

Abegglen, James C./Stalk, George, Jr., The Japanese Corporation, New York: Basic Books Inc. 1985.

Aguilar, Francis J./Yoshino, Michael Y., The Philips Group: 1987, Case No. 9-388-050, Boston, MA: Harvard Business School 1988.

Anderson, J. C./Gerbing, D. W., Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach, Psychological Bulletin, 103, 3, 1988, pp. 411-423.

Bagozzi, Richard P./Phillips, Lynn W., Representing and Testing Organizational Theories: A Holistic Construct, Administrative Science Quarterly, 27, 1982, pp. 459-489.

Bartlett, Christopher, A./Ghoshal, Sumantra, Managing Across Borders: The Transnational Solution, Boston: Harvard Business School Press 1989.

Bennigson, Lawrence A., The food Industry Jousts for Europe, Chapter 11 in Liam Fahey, (ed), Winning in the New Europe: Taking Advantage of the Single Market, Englewood Cliffs, N. J., Prentice-Hall 1992, pp. 154-165.

Bentler, Peter M., EQS: Structural Equations Program Manual, Los Angeles, CA: BMDP Statistical Software 1989.

Douglas, Susan P./Craig, C. Samuel, Examining Performance of US Multinationals in Foreign Markets, Journal of International Business Studies, Winter 1982, pp. 51-57.

Egelhoff, William G., Pattern of Control in US, UK and European Multinational Corporations, Journal of International Business Studies, Fall 1984, pp. 73-83.

Ghoshal, Sumantra, Global Strategy: An Organizing Framework, Strategic Management Journal, 8, 5, September-October 1987, pp. 425-440.

Fornell, C., A Second-generation of Multivariate Analysis: Classification of Methods and Implications for Marketing Research, in Houston, M. J. (ed.), Review of Marketing, American Marketing Association 1987, pp. 407-450.

Fornell, C./Barclay, Donald W., Jackknifing: A Supplement to Lohmoller's LVPLS Program, Ann Arbor, MI: Graduate School of Business Administration, University of Michigan 1983.

Fornell, C./Bookstein, F., A Comparative Analysis of Two Structural Equation Models: LISREL and PLS Applied to Consumer Exit-Voice Theory, Journal of Marketing Research, XIX, November 1982, pp. 440-452.

Fornell, C./Lorange, P./Roos, J., The Cooperative Venture Formation Process: A Latent Variable Structural Modeling Approach, Management Science, Special Issue on State of the Art in Theory and Methodology in Strategy Research, 36, 10, October 1990, pp. 1246- 1255.

Franko, Lawrence, G., European Business Strategies in the United States, Geneva: Business International 1971.

Hedlund, Gunnar/Nanaka, Ikujuro, Models of Knowledge Development in West and Japan, in Lorange, Peter/Chakravarthy, Bala/Ross, Johan/Van de Ven, Andrew (eds.), Implementing Strategic Process: Change, Learning and Co-operation, Oxford, England: Blackwett 1993.

Hedlund, Gunnar, The role of foreign subsidiaries in strategic decision-making in Swedish multinational corporations, Strategic Management Journal, 1, 1980, pp. 23-26.

Jaeger, A. M., The Transfer of Organizational Culture Overseas: An Approach to Control in the Multinational Corporation, Journal of International Business Studies, Fall 1983, pp. 91-114. Jeelof, Gerrit, Global Strategies of Philips, European Management Journal, 7, 1, 1989, pp. 84-91.

Johansson, Johny K./Yip, George S., Exploiting Globalization Potential: US and Japanese Strategies, Strategic Management Journal, October 1994, pp. 579-601.

Lohmoller, J., An Overview of latent Variables Path Analysis, Paper presented at the Annual Meeting of the American Educational Research Association, New York 1982.

Martinez, Jon I/Jarillo, J. Carlos, The Evaluation of Research on Coordination Mechanisms in Multinational Corporations, Journal of International Business Studies, 20, 3, Fall 1989, pp. 489-514.

Ohmae, Kenichi, Triad Power: The Coming Shape of Global Competition, New York: Free Press 1985.

Ohmae, Kenichi, The Borderless World: Power and Strategy in the Interlinked Economy, New York: Harper Business 1990.

Perlmutter, Howard V., The Tortuous Evolution of the Multinational Corporation, Columbia Journal of World Business, January-February 1969, pp. 9-18.

Porter, Michael E, Changing Patterns of International Competition, California Management Review, vol. 28, no. 2, Winter 1986, pp. 9-40.

Porter, Michael, E., The Competitive Advantage of Nations, New York: The Free Press 1990. Prahalad, C. K./Doz, Yves L., The Multinational Mission: Balancing Local Demands and Global Vision, New York: Free Press 1987.

Putti, Joseph/Singh Kulwant/Stoever, William A., Autonomy and Localization of American, European, and Japanese MNC Subsidiaries in Singapore, International Research in the Business Disciplines, vol. 1, JAI Press Inc. 1993, pp. 107-123.

Roth, Kendall/Schweiger, David/Morrison, Allen J., An Empirical Analysis of Porter's International Strategy Types, presented at the Annual Conference of the Academy of Management 1989

Stopford, John M./Wells, Louis T., Jr., Managing the Multinational Enterprise, New York: Basic Books 1972.

Tsurumi, Yoshi, Japanese and European Multinationals in America: A Case of Flexible Corporate Systems, in Macharzina, Klaus/Staehle, Wolfgang H. (eds.), European Approaches to International Management, Berlin: Walter de Gruyter 1986, pp. 23-38.

Wold, H., Systems under Indirect Observation Using PLS, in Fornell, C. (ed.), A Second Generation of Multivariate Analysis, New York: Praeger 1982, pp. 325-347.

Yip, George S., A Performance Comparison of Continental and National Businesses in Europe, International Marketing Review, 8, 2, 1991, pp. 31-39.

Yip, George S., Total Global Strategy: Managing for Worldwide Competitive Advantage, Englewood Cliffs, NJ: Prentice Hall 1992

Authors

George S. Yip, Adjunct Professor of Strategy and Organization, Anderson Graduate School of Management, University of California at Los Angeles, Los Angeles, CA, USA.

Johny K. Johansson, McCrane/Shaker Professor of International Business and Marketing, School of Business Administration, Georgetown University, Washington DC, USA.

Johan Roos, Professor, International Institute for Management Development (IMD), Lausanne, Switzerland.