Strategic Management Journal                                                                                         Strat...
638         F. Vermeulen and H. Barkemaa theoretical argument why and how the process           THE INTERNATIONAL EXPANSIO...
Internationalization, Process Dependence and Profitability                 639local knowledge and capabilities (see also He...
640         F. Vermeulen and H. BarkemaHYPOTHESES                                               negatively moderates the (...
Internationalization, Process Dependence and Profitability                 641plethora of new impressions, signals, and exp...
642         F. Vermeulen and H. Barkema                       foreign subsidiaries                       total number of  ...
Internationalization, Process Dependence and Profitability                643a relatively large number of years, sufficient ...
644          F. Vermeulen and H. BarkemaIrregularity                                                        product divers...
Internationalization, Process Dependence and Profitability                 645Analysis                                     ...
646         F. Vermeulen and H. BarkemaTests of the hypotheses                                is sampled out through the i...
Internationalization, Process Dependence and Profitability                                                 647             ...
648         F. Vermeulen and H. Barkema3), the direct effect on profitability appears negligi-   specification, we reestimat...
Internationalization, Process Dependence and Profitability                649by showing that how much an MNC benefits from  ...
650         F. Vermeulen and H. Barkemaand survival, as suggested by recent research           results are based on firms h...
Internationalization, Process Dependence and Profitability                    651process, where history and time matter, an...
652         F. Vermeulen and H. Barkema   reason why. Strategic Management Journal 10(5):          Hitt MA, Harrison JS, I...
Internationalization, Process Dependence and Profitability                    653Lebas M, Weigenstein J. 1986. Management c...
Upcoming SlideShare
Loading in …5

Pace rhythmandscope


Published on

Published in: Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Pace rhythmandscope

  1. 1. Strategic Management Journal Strat. Mgmt. J., 23: 637–653 (2002) Published online 28 March 2002 in Wiley InterScience ( DOI: 10.1002/smj.243 PACE, RHYTHM, AND SCOPE: PROCESS DEPENDENCE IN BUILDING A PROFITABLE MULTINATIONAL CORPORATION FREEK VERMEULEN1 * and HARRY BARKEMA2 1 London Business School, London, U.K. 2 Department of Organization and Strategy, Tilburg University, Tilburg, The Netherlands Many potential benefits of foreign expansion have been identified in the literature, yet empirical support that multinational firms perform better than domestic firms is mixed. This paper takes a longitudinal perspective and argues that how much a firm benefits from having foreign subsidiaries depends on its process of internationalization. We argue that a firm’s capacity to absorb expansion is subject to constraints: some expansion patterns increase profitability less than others, owing to diseconomies of time compression. We hypothesize that the speed of internationalization, the spread of the geographical and product markets entered, and the irregularity of the expansion pattern negatively moderate a firm’s increase in profitability resulting from international expansion. Model estimations based on panel data raised strong support for these predictions. Copyright  2002 John Wiley & Sons, Ltd.INTRODUCTION and Li, 1996; Hitt, Hoskisson, and Kim, 1997; Geringer, Tallman, and Olsen, 2000). Apparently,Since the early days of international business some firms manage to increase their profitabil-research, theorists have argued that firms can ity in response to international expansion whilerealize substantial benefits expanding into foreign others don’t. We expect that there are importantcountries (e.g., Hymer, 1960; Vernon, 1966). Firms contingencies regarding the relationship betweenmay, for instance, reduce market imperfections internationalization and firm profitability which atthrough internalization (Rugman, 1979, 1981), or present are insufficiently understood.realize economies of scale and scope, which allows In this paper, we focus on the process of inter-them to increase their profitability (Franko, 1989; national expansion. Although it has long been rec-Kobrin, 1991). Setting up foreign subsidiaries may ognized that organizations face constraints withalso foster innovation and knowledge transfer, respect to their growth and development (e.g.,which increases the firm’s long-term performance Penrose, 1959; Cyert and March, 1963), littleand viability (Bartlett and Ghoshal, 1989; Kogut research has directly examined how different ratesand Zander, 1992, 1993; Barkema and Vermeulen, and patterns of expansion may result in perfor-1998). However, while theorists have emphasized mance differences between firms. We approach thisthe potential gains from internationalization, the issue head-on. We investigate how the relation-empirical evidence on the impact of a firm’s ship between foreign subsidiaries and firm prof-international posture on its profitability is decid- itability is moderated by various characteristicsedly mixed (for reviews see, for instance, Tallman of its international expansion process, in terms of where and when subsidiaries were established.Key words: international expansion; development paths; Drawing from the notions of time compression dis-time compression diseconomies*Correspondence to: F. Vermeulen, London Business School, economies (Dierickx and Cool, 1989) and absorp-Regent’s Park, London, NW1 4SA, U.K. tive capacity (Cohen and Levinthal, 1990) we buildCopyright  2002 John Wiley & Sons, Ltd. Received 23 January 2001 Final revision received 14 November 2001
  2. 2. 638 F. Vermeulen and H. Barkemaa theoretical argument why and how the process THE INTERNATIONAL EXPANSIONof expansion matters. From this perspective, we PROCESS AND ITS CONSTRAINTSidentify several concrete characteristics regardinghow a firm’s profitability depends on the process Potential benefits of international expansionof growth when developing from a domestic firmto a multinational corporation (MNC). In particu- The international business and strategy literaturelar, we examine the effects of the pace, the rhythm, has suggested many reasons why substantial prof-and the geographic and product scope of a firm’s its may be realized from building an MNC. Frominternational expansion process. an economic perspective, as in industrial organiza- The process of expansion matters because build- tion models and transaction cost economics (e.g.,ing an MNC is a highly complex task (e.g., Hymer, 1960; Rugman, 1979; Caves, 1982), it hasHedlund, 1994; Malnight, 1995, 1996). Foreign been argued that MNCs benefit from increasedexpansion is constrained because the firm has market power and internalization in response toto learn how to operate in a variety of cultural market imperfections. Operations beyond domes-and institutional settings, how to set up novel tic borders enable firms to reap tax benefits, tooperations or acquire existing ones in unfamiliar benefit from common purchasing, to avoid highlocations, and how to deal with new suppliers, transaction costs, and to exploit low-cost sourcescustomers, governments, and competitors. It also of labor (Vernon, 1966; Hennart, 1982). Increasedneeds to adapt home-grown mental maps, orga- sales due to foreign expansion also allow firms tonizational structures, systems, and processes to spread R&D, marketing costs, and so on, across athe international setting (Barkema and Vermeulen, larger number of units (Franko, 1989). And, for-1998). However, since the capacity of a firm to eign direct investments imply (additional) optionsexpand and absorb new experiences is limited for MNCs to leverage their production to for-(Penrose, 1959; Cohen and Levinthal, 1990), learn- eign locations when deemed favorable (Kogut anding how to operate in a variety of foreign settings Kulatilaka, 1994).cannot endlessly be compressed in time (Dierickx Over the last decade, an increasing number ofand Cool, 1989). Therefore, even if substantial per- researchers have adopted a behavioral perspectiveformance benefits can be reaped from setting up on international expansion, and many of them havesubsidiaries abroad, it requires balanced growth to explored a second set of benefits: accruing from therealize this potential. (social) interaction between units within an MNC. Our hypotheses were tested on panel data on Examples are the benefits of learning from foreign22 firms that expanded abroad over a period of subsidiaries, and of the transfer of intangible assets25 years (1967–92). We developed measures of overseas (e.g., Ghoshal and Bartlett, 1990; Kogutthe speed of international expansion (i.e., pace), and Zander, 1992, 1993; Hedlund, 1994; Barkemathe dispersion of the internationalization process and Vermeulen, 1998). For instance, Kogut andacross different geographic and product markets Zander (1992, 1993) emphasize that building an(i.e., scope), and the regularity of the expansion MNC facilitates the knowledge transfer acrosspattern (i.e., rhythm), and tested whether these countries (i.e., within the MNC). Barkema andvariables moderate the relationship between for- Vermeulen (1998) argue that the need to adapteign subsidiaries and firm profitability. Our results to foreign settings when setting up foreign sub-corroborate a key notion of international business sidiaries may lead to temporary problems andtheory: that firms can increase their profitability local search (cf. Simon, 1959), but also to inno-due to international expansion. However, consis- vations in products, marketing, and organizationaltent with our theory, we also found that the firms practices (cf. Ghoshal, 1987; Kim, Hwang, andin our sample only realized this potential if they Burgers, 1993; Almeida, 1996), which may sub-had selected a growth strategy that was balanced sequently spread to other units within the MNCwith respect to the speed, the scope, and the regu- (Ghoshal and Bartlett, 1990; Hedlund, 1994). Inlarity of the expansion process. Thus, our research fact, recent inductive research (Birkinshaw, 1997;shows that a firm’s profitability not only depends Malnight, 1995, 1996) shows that over time for-on its (current) strategic posture, such as its inter- eign subsidiaries may obtain important roles innational diversification, but also on how it was developing, testing, and marketing new products,built. which allows MNCs to further capitalize on uniqueCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  3. 3. Internationalization, Process Dependence and Profitability 639local knowledge and capabilities (see also Hed- and absorb the complexities that accompany inter-lund, 1986; Bartlett and Ghoshal, 1989; Birkin- national expansion. Our theory, which builds onshaw and Hood, 1998; Birkinshaw, Hood, and the concepts of ‘time compression diseconomies’Jonsson, 1998). Hence, also from a behavioral per- (Dierickx and Cool, 1989) and ‘absorptive capac-spective, setting up foreign subsidiaries may imply ity’ (Cohen and Levinthal, 1990), will allow usconsiderable gains. to explain why some expansion processes imply larger benefits than others, even though the result- ing international posture may be identical.Complexities of international expansion Dierickx and Cool (1989) introduced theThe behavioral research on the international expan- concept of time compression diseconomies: thesion of firms has also emphasized the complexities fundamental mechanism of diminishing returnsof establishing and managing subsidiaries in for- when—everything else equal—the pace of processes increases. They explain it by providingeign countries. One source of complexity is that the example of MBA students in a 1-year program,firms have to learn how to operate in a vari- who may not accumulate the same stock ofety of institutional and cultural settings (Johan- knowledge as students in a 2-year program, evenson and Vahlne, 1977; Lane, 1995). In every new if all inputs other than time are doubled. We arguelocation, the firm and its management need to that the same mechanism applies to companiesinvest time and attention to establish the firm’s that establish foreign subsidiaries. Firms canpresence, hire and train a new labor force, or handle and benefit from new expansions, but theidentify a suitable acquisition candidate, and inte- amount of new experience they can absorb andgrate the new subsidiary into the MNC (David- put to commercial use (Cohen and Levinthal,son, 1983). Each new subsidiary confronts a firm 1989, 1990, 1994) is constrained in time. Newand its managers with new experiences in terms subsidiaries have to be identified, built up, andof customers, competitors, and stakeholders (e.g., integrated into the firm, but managers are boundedLi, 1995; Barkema, Bell, and Pennings, 1996). in terms of their rationality (Simon, 1959) andLearning from foreign subsidiaries and the knowl- cognitive scope (Sutcliffe, 1994). Furthermore,edge transfer within a company also requires the due to inertia, organizations are slow to adaptcareful assimilation of newly formed subsidiaries to new circumstances and configurations (Hannan(Malnight, 1996). An additional source of com- and Freeman, 1984). New structures, processes,plexity for internationalizing firms is the need and routines need to be worked out and fine-to adapt their systems, processes, and organiza- tuned over the course of time (Nelson and Winter,tional structures to the international setting (Stop- 1982). Too much foreign expansion in too short aford and Wells, 1972; Bartlett and Ghoshal, 1989). period of time leaves the firm with inappropriateFirms have ‘mental maps,’ which permeate and structures and models. Or, as Eisenhardt andunderpin their structures and processes (Perlmut- Martin (2000) put it: ‘experience that comester, 1969; Bartlett and Ghoshal, 1989; Murtha, too fast can overwhelm managers, leading to anLenway, and Bagozzi, 1998). International expan- inability to transform experience into meaningfulsion requires them to adapt these home-grown learning.’mental maps and consequently their structures, While some benefits of international expansionsystems, and processes rooted in these maps, to (e.g., tax benefits, common purchasing, tappingfit an international setting (Calori, Johnson, and into low-cost sources of labor) may be relativelySarnin, 1994a; Nohria and Ghoshal, 1994). Such easy to accomplish, other potential benefits, suchprocesses are complex and take time (e.g., Calori, as those resulting from the social interaction withinLubatkin, and Very, 1994b; Hastings, 1999; Tsai, MNCs (e.g., learning from foreign subsidiaries;2000). new roles of foreign subsidiaries to capitalize on local knowledge and capabilities; weaving newOrganizational constraints subsidiaries into the MNC) appear to be more difficult to realize. Such benefits require the carefulIn fact, we will argue that the extent to which orga- absorption of foreign ventures within the firm and,nizations are able to realize the above-described as a result, may be subject to diseconomies of timebenefits is constrained by their capacity to handle compression.Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  4. 4. 640 F. Vermeulen and H. BarkemaHYPOTHESES negatively moderates the (positive) impact of the foreign operations of an MNC on its profitability.Pace Hypothesis 1: A faster foreign expansion paceWe will argue that the contribution of foreign sub- negatively moderates the impact of a firm’s for-sidiaries to the profitability of an MNC is not eign subsidiaries on its profitability.automatic or fixed, but contingent on its develop-ment process, since there are limits to the amountof expansion the firm can absorb within a given Product scope of the expansion processperiod of time. For instance, time compression dis-economies may emerge dependent on the amount Our first hypothesis dealt with the impact of theof expansion the firm undertakes within a given amount of foreign expansion a firm undertakesperiod of time, i.e., dependent on the speed or pace in a given period of time, in terms of the num-of the internationalization process. ber of foreign subsidiaries the firm establishes. Time compression diseconomies during interna- In addition, our second hypothesis focuses on thetional expansion emerge, for one, because bounded businesses in which the expansion takes place.rationality and limited cognitive scope imply that We expect that a firm’s absorptive capacity issearch and decision making are imperfect and take not only taxed by a fast international expansiontime (Simon, 1959). Firms that expand into foreign pace, but also by the diversity of businesses thatcountries at a high pace—perhaps even with sev- are entered in the process. Diseconomies of timeeral subsidiaries at the same time—will have little compression emerge when an organization is lesstime to evaluate their foreign experience, assimi- able to absorb the expansion and the new expe-late it, and apply it to commercial ends (Cohen and riences this accompanies. This may occur dueLevinthal, 1994). Hence, when initiating new for- to the sheer number of ventures, but also dueeign expansions at a high pace, it is less likely that to their dispersion into different product markets,the firm will realize the full profit potential of these which we label the product scope of the expansionnew expansions. A high internationalization pace process.makes it more likely that (top) management of the New businesses require new knowledge, andMNC will devote suboptimal time and attention different routines and business practices. Newto carefully building greenfields, or to screening, businesses also imply that a firm and its man-selecting, and implementing acquisitions, weav- agement may have to learn a different corporateing them into the existing system of subsidiaries, culture (Reynolds, 1986; Chatman and Jehn, 1994;and carefully nurturing their role within the MNC Phillips, 1994), perhaps even another core logic(Birkinshaw, 1997; Birkinshaw and Hood, 1998). (Prahalad and Bettis, 1986). Hence, it takes con-Tsai (2000), for instance, showed that it might take siderable time and attention of a firm’s managersconsiderable time before linkages between differ- to successfully enter novel businesses, particu-ent units in a multinational company start to form. larly if the novel business is in a foreign coun-Unremitting expansion will be particularly toil- try. When foreign expansion coincides with prod-some and difficult to absorb when the MNC is uct diversification, it becomes more likely thatstill restructuring to fit its international environ- a firm makes suboptimal choices when settingment in terms of its mental maps, organizational up new subsidiaries, or when screening, select-structures, systems, and processes (Barkema and ing, and implementing acquisitions, due to causalVermeulen, 1998). Diseconomies also emerge if ambiguity and information overload (Huber, 1991;the firm has not been able yet to learn from its Haleblian and Finkelstein, 1999). Moreover, diver-prior experiences and to apply them throughout sifying into new businesses while international-the organization. izing makes organizational adaptation—in terms Hence, the more foreign subsidiaries a firm of organizational processes, processes, and sys-tries to establish in a given period of time, the tems—particularly toilsome (Barkema and Ver-more likely it will suffer from diseconomies of meulen, 1998). The simultaneous jump into (rel-time compression. This negatively affects the prof- atively) new institutional and cultural settings andits that accrue from its foreign subsidiaries. In into new businesses makes it more likely that theother words, a faster pace of foreign expansion firm is unable to fully understand and absorb theCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  5. 5. Internationalization, Process Dependence and Profitability 641plethora of new impressions, signals, and experi- adapt their structures as well (Bartlett and Ghoshal,ences. In sum, when expanding abroad, a simul- 1989). Building these systems and structures takestaneous dispersion into new businesses makes a considerable time and attention. Unless the firmfirm encounter diseconomies of time compression allows itself sufficient time, the expansion processsooner. Formally: becomes toilsome. In contrast, an expansion process that takes a Hypothesis 2: A higher product scope of the firm into a limited number of countries is much expansion process negatively moderates the more easy to digest. As a result, companies that impact of a firm’s foreign subsidiaries on its expand into just a few geographical markets suf- profitability. fer less from time compression diseconomies than firms that disperse into many markets. We expect that the larger the geographic scope of an expan-Geographic scope of the expansion process sion process, the more time an MNC needs to fullyWhile our prior hypothesis implied that dispersion absorb the accompanying experiences. Hence, theinto multiple businesses negatively influences a moderating effect of the geographic scope of thefirm’s ability to increase profitability due to foreign expansion process on the relation between foreignexpansion, a similar reasoning applies to expan- subsidiaries and company performance is negative.sion into multiple countries, i.e., the geographicscope of a firm’s expansion process. Placing ven- Hypothesis 3: A higher geographic scope oftures into many different countries is a complicated the expansion process negatively moderates theprocess compared to concentrating on a limited impact of a firm’s foreign subsidiaries on itsnumber of geographic markets. While, eventually, profitability.being active in many countries may have a positiveinfluence on a company’s performance (Geringer, RhythmBeamish, and daCosta, 1989; Barkema and Ver-meulen, 1998), the route to get there is a difficult Firms that set up foreign subsidiaries face timeone. The more countries involved in an expansion compression diseconomies because there are limitsprocess, the more difficult it becomes to absorb to their capacity to absorb expansion, in terms ofthe experience, which may lead to diseconomies the novel experiences this produces and its conse-of time compression. quences. A firm’s absorptive capacity (Cohen and Individual countries have unique features in Levinthal, 1994), however, is not necessarily con-terms of their cultural and institutional characteris- stant. For instance, it is influenced by the extenttics; local networks with suppliers and customers; to which it is utilized. Specifically, we argue thatlanguages; nature of contacts with the national overload—caused by a very high pace—reducesgovernment; education systems, and so forth (e.g., a firm’s capacity to further absorb expansion, asLane, 1995). Given a certain pace of the inter- does prolonged non-use (i.e., no expansion). Alongnationalization process, a higher geographic scope these lines, we argue that the profits that firmsimplies that the firm has to learn about more unique can realize from their foreign expansions are notnational settings, which requires more time and only influenced by the pace and dispersion of theattention than if the expansions were to take place internationalization process, but also by the (simul-in a smaller number of countries. The organization taneous) regularity of the process, or the rhythmand its management have to absorb a larger vari- at which new subsidiaries are established. We willety of experiences since the firm has to familiar- argue that firms that follow a constant, rhythmicize itself with new customers, build relationships pace are better able to benefit from foreign expan-with new suppliers, identify and understand com- sion than firms that expand in an irregular, ad hocpetitors, and so forth (e.g., Ghoshal and Bartlett, fashion.1990)—all of which taxes a firm’s absorptive Consider the example of the expansion pathscapacity. Moreover, subsidiaries in different cir- of the two firms depicted in Figure 1. Over thecumstances ask for different organizational sys- years, Firm A established its foreign subsidiariestems and processes (e.g., Lebas and Weigenstein, in a rhythmic, regular fashion; it may, for instance,1986; Gupta and Govindarajan, 1991) and com- have expanded with one subsidiary every year.panies active in a variety of cultures need to However Firm B, which at the end of the periodCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  6. 6. 642 F. Vermeulen and H. Barkema foreign subsidiaries total number of expansions per year number of time time Figure 1. Rhythmic and irregular expansion patternshas an equal number of subsidiaries, expanded absorptive capacity as well (Cohen and Levinthal,through a very different pattern: years of rapid 1990; Eisenhardt and Martin, 2000). Organizationsexpansion are alternated with long periods of gradually forget what they have learned (Bailey,inactivity. Organizations with expansion patterns 1989; Argote, Beckman, and Epple, 1990; Darr,like Firm A utilize but do not overstretch their Argote, and Epple, 1995), while they get moreabsorptive capacity. They are able to interpret and rigidly locked into their existing structures, sys-absorb their experiences because they can relate tems, cultures, and mental models (Lewin, 1936;them to similar actions in their recent past (Ellis, Miller, 1993; Bettis and Prahalad, 1995). Owing1965; Cohen and Levinthal, 1989; Barkema et al., to the lower absorptive capacity, firms following1997). As a result, they are flexible and capa- an irregular expansion path will encounter timeble to implement and absorb additional expansions compression diseconomies sooner than firms that(Hitt et al., 1998; Vermeulen and Barkema, 2001). expand in a rhythmic pattern. Thus, we expectFirms that implement and absorb changes in an that in internationalizing firms where managementoptimal, rhythmical pattern may even reach a state has ‘rhythmically’ initiated foreign expansion, sub-of ‘flow,’ as Brown and Eisenhardt (1997) found sidiaries contribute more to profitability than infor firms adopting innovations in the computer companies that established their foreign presenceindustry. in an irregular, ad hoc fashion. Patterns as displayed by Firm B, however,involve large peaks of rapid expansion followed by Hypothesis 4: An irregular pace negatively mod-long periods of inactivity. Peaks of rapid expansion erates the impact of a firm’s foreign subsidiariesmay lead to overload, and organizations and man- on its profitabilityagers that experience overload see their absorp-tive capacity reduced (Simon, 1959; Huber, 1991).They are unable to further absorb expansions METHODOLOGYbecause they are unable to interpret and assess Datathem, while being left with systems and struc-tures unfit to accommodate (additional) foreign To test the hypotheses, we collected a longitu-subsidiaries. Periods of inactivity reduce a firm’s dinal database on multinational firms, coveringCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  7. 7. Internationalization, Process Dependence and Profitability 643a relatively large number of years, sufficient to Venkatraman and Ramanujam, 1986; Hitt et al.,determine their patterns of internationalization. For 1997). The models shown below are based on apractical reasons, we did not select a random sam- 3-year moving average of return on assets, sinceple, but examined a set of firms that had existed firms are (to some extent) able to manipulate inover a longer period of time. which year they report profits or losses. Models We started with all firms in the main segment based on 1-year observations of return on assetsof the Amsterdam Stock Exchange—a total of led, however, to virtually identical results.40 firms—in 1992, and excluded all financialinstitutions. All remaining companies were Dutch(with six firms of the final sample cross-listed). Foreign subsidiariesNext we determined the timing and location of Foreign subsidiaries were measured yearly as thetheir foreign ventures. Data were mainly acquired number of foreign affiliates of the firm. Affili-from annual reports, but in later stages most firms ates had to represent a ‘physical’ foreign presence,were also contacted to verify and, sometimes, ranging from a sales office to production complete information. We traced these firms By the end of our sample period (i.e., the end ofback in time until we had a sufficient number of 1992) the average firm in the sample had estab-years available. This turned out to be in the year lished 34 subsidiaries.1967—further back in time annual reports becamevery concise, and/or difficult to obtain. Severalfirms had to be eliminated from the sample because Speedthey came into existence and/or became listedon the Stock Exchange well after 1967. Others ‘Speed’ indicates how many foreign expansions awere eliminated from the sample because they firm undertakes in a certain period of time. There-appeared to have internationalized earlier—well fore, to measure speed, the average number of for-before the year 1967. This resulted in a panel of eign subsidiaries per year was computed, i.e., the572 observations: 22 firms covering 26 years. number of foreign subsidiaries divided by the num- The firms in the sample are fairly large compa- ber of years since the firm’s first foreign expansion.nies—their average sales were about 2.5 billion A large (average) number of expansions per yearNLG with 14,000 employees—with an average indicates a fast-paced international expansion pro-percentage of sales abroad of 63 percent (median: cess. Alternatively, speed can be measured through63%) at the end of the observation period. In the variable ‘number of years since the firm’s firsttotal, the firms undertook 741 foreign expan- foreign expansion,’ i.e., how many years it took thesions, of which 67 percent were within the EU. firm to reach its current international posture. ThisThey were active in a wide variety of indus- alternative variable—based on the same informa-tries, including manufacturing office equipment, tion—led to identical results.precision machinery, paper and packaging, foodproducts, pharmaceutical and chemical products, Geographic scopebrewing, publishing and printing, retailing, trading,and tank storage. Obviously, however, it is not a The number of countries in which a firm estab-random sample. They are all firms that internation- lished subsidiaries during its international expan-alized throughout the last few decades. Moreover, sion is used to measure the geographic scope ofcollecting data working back in time creates a bias the expansion process. At the end of the sampletowards surviving firms. How internationalization period, the average firm had expanded into 10 dif-and the pattern of expansion influences the survival ferent countries.chances of firms is a topic well worth examining,but beyond the scope of the current paper. Product scopeVariables The product scope of a firm’s expansion process is measured as the number of additional 3-digit SBIProfitability codes entered by the firm while internationalizing.The dependent variable is firm profitability. It was The SBI code is the Dutch equivalent of themeasured through the firm’s return on assets (e.g., SIC code.Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  8. 8. 644 F. Vermeulen and H. BarkemaIrregularity product diversity of the firm, that is, the number of businesses—measured as 3-digit SBI codes—inRegularity, or rhythm, of the internationalizationprocess was measured through the kurtosis of the which the firm was active during the expansion.first derivative of the number of foreign ventures Product diversity may influence both internation-of the firm over time. This variable measures how alization (Fouraker and Stopford, 1968; Hitt et al.,concentrated in time the change in the number 1994) and firm performance (Hoskisson and Hitt,of foreign subsidiaries is. Figure 1 illustrates this 1990; Datta, Rajagopalan, and Rasheed, 1991).relationship. The upper graphs depict the level of Consistent with prior research (Tallman and Li,internationalization of the firm, i.e., the number 1996; Hitt, Hoskisson, and Ireland, 1997), weof foreign subsidiaries. The bottom graphs depict added the square of this variable to capture non-the change (i.e., the first derivative) in internation- linearities.alization. How concentrated in time the changes Second, we controlled for firm size through thein internationalization are is measured through the firm’s assets, corrected using a price index. Again,kurtosis of this distribution: the square was added to control for nonlinearities, n(n + 1) as found in previous research (Haveman, 1993a). kurtosis = Larger firms may benefit from economies of scale (n − 1)(n − 2)(n − 3) or scope (Franko, 1989), while very large firms xi − x 4 3(n − 1)2 become rigid and inert (Hannan and Freeman, − s (n − 2)(n − 3) 1977). Following prior research (Hitt et al., 1997), wewhere n = number of observations, xi = number also controlled for the firms’ financial structureof expansions in year i, and s = standard deviation through a debt ratio (total liabilities to assets),of the number of expansions. since capital structure may affect a firm’s ability Large peaks in a firm’s expansion pattern, com- to expand, as well as its performance (Jensen,bined with periods of inactivity, result in a rel- 1986).atively concentrated distribution and therefore a Finally, following prior research (Hitt et al.,high kurtosis. A constant pace of foreign expan- 1997), we also controlled for the number of foreignsion—that is, a rhythmic or regular expansionpattern—results in a relatively flat distribution and acquisitions and the number of equity alliancestherefore a lower kurtosis. during each year, since these modes of expansion may be related to both internationalization and performance (Gulati, 1995; Barkema andControl variables Vermeulen, 1998).A number of control variables were included in Table 1 displays summary statistics and partialthe analyses. First, we controlled for the level of correlations of the variables.Table 1. Means, standard deviations, and partial correlationsaVariables Means S.D. 1 2 3 4 5 6 7 8 9 1. Return on assets 6.98 5.76 2. Foreign subsidiaries 13.9% 15.0% 0.03 3. Speed 0.94 0.78 −0.05 0.53 4. Product diversity 7.38 5.29 −0.08 0.08 0.08 5. Number of countries 6.34 4.70 −0.00 0.55 0.53 0.03 6. Irregularity 2.26 3.36 −0.00 0.02 −0.01 −0.08 −0.24 7. Firm size 5.03 7.78 0.09 0.11 0.20 0.24 −0.06 −0.08 8. Debt ratio 0.42 0.15 −0.26 0.20 0.20 0.23 0.15 −0.03 0.09 9. Acquisitions 1.73 2.17 0.10 0.19 0.25 0.17 0.09 0.01 −0.01 0.0010. Equity alliances 0.11 0.36 0.06 −0.01 0.05 0.07 −0.02 −0.04 −0.07 0.00 0.10N = 572. Correlations with absolute value greater than 0.08 are significant at the 5 percent level.a Pearson correlations partialled for fixed firm-effects and a free time polynoom.Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  9. 9. Internationalization, Process Dependence and Profitability 645Analysis since speed may not have a direct effect on firm performance, but only as a moderator. LikewiseWe estimated ordinary least squares, fixed-effects for the interactions with product scope, geographicmodels since the assumptions necessary to employ scope, and process irregularity.random effects would be violated in the current Finally, to make sure that significance of theseresearch setting (Hsiao, 1986), as confirmed by interactions was not caused by ‘spurious correla-a Hausman test. Hence, we used firm dummies tion’ (e.g., Aiken and West, 1991; Ganzach, 1998),to control for firm-specific effects. A polynoom, we also tested each of the interactions in mod-consisting of calendar time and several of its els that included squares of the main variables.powers, was included to control for any possi- These analyses led to identical conclusions to thoseble (possibly nonlinear) development over time reported below.throughout the sample period. However, modelswith yearly time dummies led to virtually identicalresults. Since our hypotheses concern moderating RESULTSeffects, we used interactions to test them. Wepredicted that the speed, the product scope, the Table 2 provides the statistical results of thegeographic scope, and the irregularity of the hypotheses tests. Model 1 shows the estimatesexpansion process will all reduce the beneficial without interactions. The effect of the variableeffect of foreign subsidiaries on firm profitability. ‘number of foreign subsidiaries’ is positive andTherefore we expect that, for instance, the significant, showing that, on average, the firms ininteraction between foreign subsidiaries and speed our sample experienced an increase in profitabilitywill be negative. The interpretation is that foreign due to their international expansion. The size of thesubsidiaries—i.e., the main term—may have a coefficient implies that, for instance, firms with 10positive influence on firm profitability, but that foreign subsidiaries have a return on assets whichthe benefits will be smaller when the speed of is, on average, 1.08 percent higher than when theyinternational expansion is high. The estimate on were purely domestic firms. Interaction terms werethe main term of speed itself may be insignificant entered into Models 2–5 successively. Model 6(Jaccard et al., 1990; Aiken and West, 1991), shows the results of the full model.Table 2. Regression of firm profitability Model 1 Model 2 Model 3 Model 4 Model 5 Model 6Tests of the hypothesesForeign subsidiaries × speed −0.108∗∗∗ −0.066∗Foreign subsidiaries × product scope −0.009∗ −0.011∗Foreign subsidiaries × geographic scope −0.017∗∗∗ −0.009∗Foreign subsidiaries × irregularity −0.005∗ −0.009∗∗Control variablesIntercept 9.914∗∗∗ 10.66∗∗∗ 10.16∗∗∗ 10.24∗∗∗ 10.03∗∗∗ 11.04∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗Foreign subsidiaries 0.108 0.412 0.159 0.357 0.133 0.539∗∗∗Speed −0.641 −0.169 −0.678 −1.827 −0.806 −1.340Irregularity −0.048 −0.009 −0.065 0.017 0.033 0.127Geographic scope −0.309∗∗ −0.363∗∗∗ −0.340∗∗∗ 0.096 −0.319 −0.169Product diversity 0.359 0.464∗ 0.506∗∗ 0.359 0.338 0.564∗Product diversity squared −0.020∗∗ −0.025∗∗∗ −0.019∗∗∗ −0.021∗∗ −0.019∗∗ −0.022∗∗∗Firm size 0.459∗∗ 0.494∗∗∗ 0.470∗∗∗ 0.563∗∗∗ 0.488∗∗ 0.601∗∗∗Firm size squared −0.005∗ −0.006∗ −0.006∗∗ −0.007∗∗ −0.006∗ −0.008∗∗Debt ratio −10.23∗∗∗ −11.15∗∗∗ −10.53∗∗∗ −11.04∗∗∗ −10.15∗∗∗ −11.49∗∗∗Acquisitions 0.204∗ 0.204∗ 0.200∗ 0.193∗ 0.186∗ 0.163Equity alliances 0.791 0.717 0.835 0.668 0.768 0.692R2 0.64 0.65 0.64 0.65 0.64 0.66∗ ∗∗ ∗∗∗ p < 0.05; p < 0.01; p < 0.001Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  10. 10. 646 F. Vermeulen and H. BarkemaTests of the hypotheses is sampled out through the inclusion of the interac- tion, the main term on geographic scope becomesHypothesis 1 predicts that the speed of internation- insignificant. This suggests that the benefits ofalization negatively moderates the relation between international expansion stem from operating in aa firm’s foreign subsidiaries and profitability. Mod- number of foreign subsidiaries abroad, rather thanels 2 and 6 display the models that include the from diversifying into different countries. We willinteraction between the variables ‘number of for- return to this issue in more depth below, where weeign subsidiaries’ and ‘speed,’ which is used to will present the results of further analysis.test this hypothesis. The estimates on foreign sub- Finally, Models 5 and 6 show that the interactionsidiaries are positive and significant, which indi- between foreign subsidiaries and the irregularity ofcates that these have a beneficial influence on the expansion process is consistently negative andthe profitability of a firm. The coefficient on the significant. This supports Hypothesis 4: firms thatmain term of the variable speed is insignificant, have expanded through an irregular pattern benefitwhich indicates that speed of international expan- less, in terms of profitability, from having foreignsion has no direct effect on profitability. However, predicted, speed does negatively moderate therelationship between foreign subsidiaries and prof-itability. Hence, the models show that the sub- Size of the effectssidiaries of companies that very quickly expand The coefficient of the variable foreign subsidiariesabroad contribute less to a firm’s profitability than in Model 1 implies that, on average, every addi-the subsidiaries of companies that expand more tional subsidiary increases a firm’s return on assetsslowly. This corroborates Hypothesis 1. by 0.108 percent. This number increases to 0.412 Models 3 and 6 include the interaction between percent if controlled for the negative, moderat-foreign subsidiaries and the product scope of ing effect of speed (Model 2). For example, ifthe expansion process. Hypothesis 2 implies that a firm establishes 10 subsidiaries in 5 years itsimultaneous entry into a variety of businesses would increase its return on assets by 1.96 percentnegatively moderates the (positive) impact of a (= 0.412 ∗ 10 − 0.108 ∗ 10 ∗ 10/5). Likewise, thefirm’s foreign subsidiaries on its profitability. The coefficients in Model 3 imply that a company thatestimate is consistently negative and significant, places 10 subsidiaries in seven countries increaseswhich supports Hypothesis 2. As evidenced by the its return on assets by 2.38 percent (= 0.357 ∗main term of foreign subsidiaries combined with 10 − 0.017 ∗ 10 ∗ 7). Similar computations can beits interaction with product scope, firms benefit made for the effect of product scope (Model 4) andfrom international expansion, but less so if it also irregularity (Model 5).takes them into many different businesses. These effects highlight the trade-offs in the mod- Likewise, Hypothesis 3 predicts a negative inter- els, for instance, between the positive effect of hav-action between foreign subsidiaries and geographic ing foreign subsidiaries and the moderating nega-scope of the international expansion process. The tive influence of the speed at which they are estab-interaction in Models 4 and 6 is indeed negative lished. Figure 2 illustrates this trade-off based onand significant, which corroborates the hypoth- the estimates in Table 2. Slow foreign expansionesis: the larger the number of countries where (speed = 0.5) limits the negative effects of speed,the foreign expansion takes place, the smaller the and firms following this expansion pattern bene-increase in profitability due to setting up foreign fit more from their foreign subsidiaries than firmssubsidiaries. Interestingly, in the models without that establish subsidiaries at a higher pace (speedthe interaction, the coefficient on the main term = 2 or 5), as shown by the higher, positive slope of‘geographic scope’ is negative, suggesting that the relationship depicted in Figure 2(a). Obviously,firm profitability is hampered by being active in slow speed also leaves the firm with very few sub-a number of different countries. Adding the inter- sidiaries to benefit from. Plotting the profitabilityaction reveals why: foreign subsidiaries contribute of these internationalizing companies against timeless to firm profitability when they are placed into (Figure 2b) shows that the company expanding atmany different countries. They contribute more to speed = 2 is better off than the firm expandingprofitability when they are concentrated in a lim- at speed = 0.5. It simply has more subsidiaries toited number of geographical markets. If this effect benefit from, which compensates for the negativeCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  11. 11. Internationalization, Process Dependence and Profitability 647 25 25 20 20 speed = .5 speed = 2 return on assets return on assets 15 15 speed = 2 10 speed = .5 10 5 5 speed = 5 0 66 71 76 81 86 91 year 0 −5 0 5 10 15 20 25 30 35 40 speed = 5 −10 number of foreign subsidiariesFigure 2. (a) Estimated relationship between foreign subsidiaries and firm profitability moderated by speed. (b) Estimated profitability of internationalizing firms during the course of timeeffect of speed. For instance, after 10 years it has percent (0.539 − 0.108) of the effect on firm prof-20 companies while the firm expanding at speed = itability per foreign subsidiary.0.5 has only five. Furthermore, very rapid expan-sion (speed = 5) annuls the entire beneficial effectof foreign subsidiaries; in this case international Geographic scope versus number of foreignexpansion even decreases firm profitability. subsidiaries The model estimates, however, imply that the The degree of internationalization of a firm can befirms in our sample typically did not overstrain measured in several ways, for instance, through thethemselves in terms of speed, product and size of its foreign operations (e.g., the number ofgeographic scope, and irregularity. For instance, foreign establishments) or the scope of these oper-additional computations showed that firms only ations (e.g., the number of foreign countries wherefailed to increase profitability as a result of the firm is operating). Large size does not neces-their foreign expansion if they selected expansion sarily imply large scope: firms may either choosestrategies implying that they were at least one to spread their operations over a large number ofstandard deviation above average on all four countries, or concentrate their foreign presence inmoderators. a few selected countries. Previous research has The total amount of variance explained by the typically focused on capturing either the size ordifferent models ranges from 64 percent to 66 the scope of foreign operations (e.g., Grant, Jam-percent. Hence, allowing for interactive effects in mine, and Thomas, 1988; Geringer et al., 1989;models does not explain much additional vari- Tallman and Li, 1996), or combined them into oneance; instead it reveals the different components variable, through an entropy measure (e.g., Hittthat constitute it (cf. Jaccard et al., 1990; Aiken et al., 1997). In our study we measured them sepa-and West, 1991), which enables a more fine-tuned rately: the number of foreign subsidiaries relates toview on the relationships between the different the size of foreign operations, while we measuredvariables. A good indication of the impact of the geographic scope through the number of countriesinteraction effects is the change in the main term involved. Our results clearly show that the two‘foreign subsidiaries’ induced by inclusion of inter- may have very different effects—something thatactions. In Model 1 the coefficient of this vari- is concealed when they are combined into oneable indicates that the profitability of a firm, on measure. The number of foreign subsidiaries con-average, increases by 0.108 percent per foreign sistently has a positive influence on firm profitabil-expansion. When controlling for the moderating ity throughout the different models, while in theeffects of the four process characteristics, this coef- model without interactions (Model 1) geographicficient increases to 0.539 percent. Hence, together, scope is even negative. If controlled for the mod-the four moderating variables account for 0.431 erating effect of geographic scope (cf. HypothesisCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  12. 12. 648 F. Vermeulen and H. Barkema3), the direct effect on profitability appears negligi- specification, we reestimated all models using cen-ble. Hence, internationalizing firms seem to benefit tered variables. All above-mentioned results werepredominantly from building an organization that clearly replicated.consists of multiple foreign subsidiaries. To further examine the robustness of our find- Our theory suggests that expansions dispersed ings, we also estimated models with return onacross many countries tax a firm’s absorptive assets replaced by return on equity. Although itcapacity more than expansion into a limited num- has been argued that return on assets is preferableber of geographic markets. Our arguments also to return on equity because the latter is sensitive toseem to imply that expansions into different coun- differences in capital structure (Hitt et al., 1997),tries are more easy to absorb if the countries are the estimation results were largely similar, exceptrelated (e.g., Ronen and Shenkar, 1985). Therefore, for the influence of process irregularity, whichwe also computed an entropy measure of geo- lost some of its statistical significance (althoughgraphical ‘diversification’ (Kim, 1989; Vachani, still p < 0.10). Models using a composite mea-1991) to replace our measure of geographic scope, sure based on both return on assets and return onbecause the former measure also takes into account equity led to similar conclusions to those reportedto what extent countries are alike. We defined regions according to Ronen and Shenkar Alternative entropy measures for product(1985) and assigned weights using the number diversification (Jacquemin and Berry, 1979;of subsidiaries per region, rather than sales per Hoskisson et al., 1993), based on the SBI codes inregion (Barkema and Vermeulen, 1998), to avoid which the foreign subsidiaries were active, also ledproblems with assigning exports and for reasons to highly similar findings, including a significantof data availability. Analyses with this alternative interaction with number of foreign subsidiariesmeasure replicated the results as reported above, (−0.14; p < 0.0001). Dispersion across differentwith the interaction between the number of foreign businesses appears to be particularly complicated ifsubsidiaries and geographical diversification even those businesses are very different. These findingsmore significant (−0.44; p < 0.0001). Apparently, raise further support for the perspective outlined inthe moderating negative effects of the geographic this study.scope of the expansion process are aggravated if Finally, we estimated models with lagged inde-the countries involved are highly dissimilar. This pendent variables (1, 2, or 3 years) vis-` -vis firm araises further support for our theory. Moreover, profitability, since investment in foreign operationsthis more fine-grained analysis of geographic scope may precede profit gains. The models producedresulted in a positive estimate of the direct effect results comparable to those reported above. Onlyof geographic scope on firm profitability (3.09; the estimates on product scope appeared some-p < 0.0001), in addition to the influence of the what less significant, especially in the models withnumber of foreign subsidiaries (0.81; p < 0.0001). longer time-lags, but further analysis indicatedThese results show that the process of expanding that this was mainly due to the smaller num-into very different countries is difficult and toil- ber of observations on which the estimates weresome; however, once a firm has established this based—we had to delete a number of observa-position, it may experience a positive influence tions, dependent on the specific time-lag in thefrom its dispersed foreign presence on its prof- model, due to missing values in the early yearsitability. of the sample.Sensitivity analysis DISCUSSIONWhen estimating models with interaction terms itis often advisable to center the variables involved The internationalization of a firm is a complex(Jaccard et al., 1990; Aiken and West, 1991): to task. Although many potential benefits of interna-reduce problems of multicollinearity, or when the tional expansion have been identified in the litera-variables do not have a meaningful interpretation ture, the empirical support for a key assumptionat value zero. Neither condition appears to apply of this theory—that firms successfully increaseto our models; however, to examine the sensitivity their profitability as a result of international expan-of the support from our models to this alternative sion—is mixed. Our study adds to this literatureCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  13. 13. Internationalization, Process Dependence and Profitability 649by showing that how much an MNC benefits from their efforts and benefit from their position (cf.,where it stands ‘today,’ in terms of its interna- Chang, 1995), while other growth paths may betional posture (e.g., Tallman and Li, 1996; Hitt more difficult to digest and inhibit potential bene-et al., 1997), depends on how it has arrived there. fits from materializing. From this perspective, ourWe developed and tested a theory regarding how study adds to more general theory on the growthvarious characteristics of the internationalization of firms (Penrose, 1959).process of a firm influence how much its foreign Several fundamental managerial implicationssubsidiaries ‘currently’ contribute to its profitabil- follow from our study too. Prior research hasity. Consistent with predictions, we found that the shown that high growth in one business placesspeed at which subsidiaries were established, the constraints on the ability of the firm to growdispersion of the expansion process into different in other dimensions (Galunic and Eisenhardt,countries and businesses, and the irregularity of the 1996). Our study implies that adding complexityprocess (i.e., large expansion peaks and periods of in one dimension, for instance by increasinginactivity, as opposed to a rhythmic pace) nega- the pace of the internationalization process (intively moderate how much a firm benefits from its terms of the number of foreign expansions),international operations. implies that a firm needs to restrict complexity We explained this phenomenon by arguing that in other dimensions, for instance by limiting‘diseconomies of time compression’ (Dierickx and the number of novel countries or businessesCool, 1989) emerge during the process of inter- entered during the expansion. Likewise, firmsnational expansion. We argued that these dis- that feel the need to quickly enter a relativelyeconomies exist because the capacity of a firm to large number of countries—to capture marketabsorb expansion is constrained, owing to proper- share early, obtain a global presence, or establishties such as bounded rationality, cognitive limita- a global standard (e.g., Caves, 1982)—shouldtions, and structural inertia. Hence, in our theory, be aware of the restrictions this imposes onabsorptive capacity (Cohen and Levinthal, 1990) the pace of its international expansion indrives time compression diseconomies. Interna- terms of the number of new subsidiaries thattionalizing firms that overstrain their absorptive can be successfully established and absorbed.capacity are more likely to devote suboptimal time Similar considerations apply to entering differentand attention to setting up greenfield operations; businesses and following an irregular expansionto screening, selecting, and implementing acquisi- pattern. In other words, firms need to follow a pathtions; to weaving new subsidiaries into the internal of balanced growth; they need to be aware of thenetwork, and to nurturing their role within the different trade-offs that exist and of the necessityMNC. It is also less likely that such firms have dis- to make clear strategic choices about which of thecarded obsolete home-grown mental maps, struc- different dimensions (i.e., pace, rhythm, and scope)tures, systems, and processes that are no longer they will prioritize. Otherwise they may jeopardizeoptimal, and replaced them with new ones that fit the profitability of the MNC they are building.the international environment better. As a result, Practice shows that firms are frequently cajoledthe contribution of expansions formed during such into a strategy of fast growth, because investorsa process of ‘forced’ internationalization to firm and analysts repeatedly expect commandingprofitability is limited. figures or because they imitate competitors that Our study clearly adds to prior research that make similar moves (DiMaggio and Powell, 1983;compares the different strategic postures of firms Haveman, 1993b). Bandwagon effects may evenat a given point in time. Even if two firms result in waves of mergers and acquisitions, ashave ended up in the same strategic position—for observed by prior research at different pointsinstance, having the same level of internationaliza- in time (Shleifer and Vishny, 1991; Stearnstion—they may have gone through very different and Allan, 1996). While such behavior may beexpansion processes, which may have resulted in understandable, our study suggests that it maydifferent profitability levels. Hence, to fully under- not be optimal from a strategic perspective. Yes,stand profitability differences of internationalized acquisitions may benefit a firm, since they mayfirms, one may have to look back in time and have long-term advantages that transcend thetake into account how they have arrived there. focal acquisition, by ‘revitalizing’ the firm andSome expansion routes may allow firms to absorb consequently enhancing its long-term profitabilityCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  14. 14. 650 F. Vermeulen and H. Barkemaand survival, as suggested by recent research results are based on firms headquartered in a single(Vermeulen and Barkema, 2001). However, there country. Furthermore, these firms all started ven-are constraints on how much expansion a turing abroad in the late 1960s, early 1970s, with afirm can digest, as indicated by the present large proportion of expansions taking place withinstudy. Periods of high expansion activity (for the EU. Although countries within the EU can alsoinstance, during acquisition waves) followed by be very different (Ronen and Shenkar, 1985; Lane,relative inactivity—the very definition of an 1995), additional empirical research using samplesirregular expansion pattern in our study—reduce from other countries, addressing industry-specificprofitability. A regular expansion pattern rather effects, and/or examining the influence of companythan uncontrolled herd behavior helps to build a size and other characteristics would clearly add toprofitable MNC. the current study. Diseconomies of time compres- sion may be more or less prevalent under differentLimitations and suggestions for further circumstances.research Along the same lines, absorptive capacity and, as a result, the incidence of time compression dis-Our paper used a behavioral perspective on the economies may vary with organizational character-benefits of building a multinational corporation. istics. For instance, it may depend on characteris-This also informs the boundaries of our the- tics of the top management team (e.g., Eisenhardtory. Time compression diseconomies and the phe- and Schoonhoven, 1990), or the current structurenomenon of process dependence predominantly of the firm (Barkema and Vermeulen, 1998). Futureaffect the potential benefits of foreign expansion research uncovering how these and other factorsthat result from the social interaction between moderate the impact of process pace, scope, andunits within an MNC (e.g., mutual learning, trans- regularity on the profitability of a firm’s (interna-fer of intangible assets). Our theory probably tional) expansion would complement the currentapplies much less to the potential advantages that study as well.result from higher efficiency in MNCs (e.g., com- Finally, some of the potential benefits of interna-mon purchasing, tapping into low-cost sources of tionalization identified in the literature seem morelabor)—these benefits could perhaps be achieved closely associated with the size of foreign oper-regardless of, for instance, the speed of the inter- ations (e.g., Franko, 1989; Kobrin, 1991), oth-nationalization process. ers with having operations in different countries A limitation of our study in terms of data is that (e.g., Kogut and Kulatilaka, 1994). Our study dis-the companies in our sample do not form a ran- tinguished between the number of foreign sub-dom sample. They are all firms that ‘survived’ for sidiaries of a firm and the geographic scope ofa sustained period of time (i.e., over 25 years). As its expansion. Future research aimed at furtherreported in the Results section, most of the firms in disentangling these different effects would greatlyour sample did not seem to overstrain their capac- advance our understanding of the benefits of inter-ity in terms of their speed of foreign expansion, nationalization and their contingencies.scope, and pattern regularity. This may be due tothe ‘survival bias’ in our sample: firms with highlevels on these dimensions are more likely to have CONCLUSIONfailed, i.e., to go bankrupt, or perhaps have beentaken over, and as a result were not included in Theory and practice suggest that firms may profitour study. We welcome future research examining substantially from having an international pres-the influence of expansion process characteristics ence. However, as our longitudinal study showed,examined in our study on the performance of a there are limits to how much expansion an orga-broader category of organizations, perhaps using nization can cope with. Internationalization cannotother performance measures such as the exit rate of be forced, in terms of the pace of setting up foreigncompanies (e.g., going bankrupt or not, or having subsidiaries, and the number of geographic andbeen taken over), and so on. product markets covered in the process. Irregular, Another concern is the issue of generalizabil- ad hoc growth further complicates the absorptionity. Although our firms stem from a wide variety of the foreign expansion by the MNC. We con-of industries, a limitation of our study is that the ceptualized organizational expansion as a dynamicCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  15. 15. Internationalization, Process Dependence and Profitability 651process, where history and time matter, and the evolution in relentlessly shifting organizations.path taken determines the lucre at arrival. We hope Administrative Science Quarterly 42: 1–34. Calori R, Johnson G, Sarnin P. 1994a. CEO’s cognitivethat this perspective will prove useful for future maps and the scope of the organization. Strategicinvestigation of the growth of firms and their pros- Management Journal 15(6): 437–457.perity. Calori R, Lubatkin M, Very P. 1994b. Control mecha- nisms in cross-border acquisitions: an international comparison. Organization Studies 15: 361–379.ACKNOWLEDGEMENTS Caves RE. 1982. Multinational Enterprise and Economic Analysis. Cambridge University Press: Cambridge, UK.This research was supported by the Netherlands Chang SJ. 1995. International expansion strategy ofOrganization for Scientific Research. We thank Japanese firms: capability building through sequen-John McGee, two anonymous reviewers, and semi- tial entry. Academy of Management Journal 38:nar participants at the London Business School for 383–407. Chatman JA, Jehn KA. 1994. Assessing the relationshiptheir helpful comments on this paper. between industry characteristics and organizational culture: how different can you be? Academy of Management Journal 37: 522–553.REFERENCES Cohen WM, Levinthal DA. 1989. Innovation and learn- ing: the two faces of R&D. Economic Journal 99:Aiken LS, West SG. 1991. Multiple Regression: Testing 569–596. and Interpreting Interactions. Sage: Thousand Oaks, Cohen WM, Levinthal DA. 1990. Absorptive capacity: CA. a new perspective on learning and innovation.Almeida P. 1996. Knowledge sourcing by foreign Administrative Science Quarterly 35: 128–152. multinationals: patent citation analysis in the Cohen WM, Levinthal DA. 1994. Fortune favors the U.S. semiconductor industry. Strategic Management prepared firm. Management Science 40: 227–251. Journal Winter Special Issue 17: 155–165. Cyert RM, March JG. 1963. A Behavioral Theory of theArgote L, Beckman SL, Epple D. 1990. The persistence Firm. Prentice-Hall: Englewood Cliffs, NJ. and transfer of learning in industrial settings. Darr ED, Argote L, Epple D. 1995. The acquisition, Management Science 36: 140–154. transfer and depreciation of knowledge in serviceBailey CD. 1989. Forgetting and the learning curve: a organizations: productivity in franchises. Management laboratory study. Management Science 35: 340–352. Science 42: 1750–1762.Barkema HG, Bell JHJ, Pennings JM. 1996. Foreign Datta DK, Rajagopalan N, Rasheed AMA. 1991. Diver- entry, cultural barriers, and learning. Strategic sification and performance: critical review and Management Journal 17(2): 151–166. future directions. Journal of Management Studies 28:Barkema HG, Shenkar O, Vermeulen F, Bell JHJ. 1997. 529–558. Working abroad, working with others: how firms learn Davidson WH. 1983. Market similarity and market to operate international joint ventures. Academy of selection: implications for international marketing Management Journal 40: 426–442. strategy. Journal of Business Research 11: 439–456.Barkema HG, Vermeulen F. 1998. International expan- Dierickx I, Cool K. 1989. Asset stock accumulation and sion through start-up or acquisition: a learning per- sustainability of competitive advantage. Management spective. Academy of Management Journal 41: 7–26. Science 35: 1504–1514.Bartlett CA, Ghoshal S. 1989. Managing across Borders: DiMaggio PJ, Powell WW. 1983. The iron cage revisited: The Transnational Solution. Harvard Business School institutional isomorphism and collective rationality in Press: Boston, MA. organizational fields. American Journal of SociologyBettis RA, Prahalad CK. 1995. The dominant logic: 48: 147–160. retrospective and extension. Strategic Management Eisenhardt KM, Martin JA. 2000. Dynamic capabilities: Journal 16(1): 5–14. what are they? Strategic Management Journal , SpecialBirkinshaw J. 1997. Entrepreneurship in multinational Issue 21(10–11): 1105–1122. corporations: the characteristics of subsidiary initia- Eisenhardt KM, Schoonhoven CB. 1990. Organizational tives. Strategic Management Journal 18(2): 207–229. growth: linking founding team, strategy, environment,Birkinshaw J, Hood N. 1998. Multinational subsidiary and growth among U.S. semiconductor ventures, evolution: capability and charter change in foreign- 1978–1988. Administrative Science Quarterly 35: owned subsidiary companies. Academy of Manage- 504–529. ment Review 23: 773–795. Ellis HC. 1965. The Transfer of Learning. Macmillan:Birkinshaw J, Hood N, Jonsson S. 1998. Building firm- New York. specific advantages in multinational corporations: the Fouraker LE, Stopford JM. 1968. Organizational struc- role of subsidiary initiative. Strategic Management ture and the multinational strategy. Administrative Sci- Journal 19(3): 221–242. ence Quarterly 13: 47–64.Brown SL, Eisenhardt KM. 1997. The art of continuous Franko LG. 1989. Global corporate competition: who’s change: linking complexity theory and time-paced winning, who’s losing, and the R&D factor as oneCopyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  16. 16. 652 F. Vermeulen and H. Barkema reason why. Strategic Management Journal 10(5): Hitt MA, Harrison JS, Ireland RD, Best A. 1998. At- 449–474. tributes of successful and unsuccessful acquisitions ofGalunic DC, Eisenhardt KM. 1996. The evolution of U.S. firms. British Journal of Management 2: 91–114. intracorporate domains: divisional charter losses Hitt MA, Hoskisson RE, Ireland RD. 1994. A mid-range in high-technology, multidivisional corporations. theory of the interactive effects of international and Organization Science 7: 255–282. product diversification on innovation and performance.Ganzach Y. 1998. Nonlinearity, multicollinearity and the Journal of Management 20: 297–326. probability of type II error in detecting interaction. Hitt MA, Hoskisson RE, Kim H. 1997. International Journal of Management 24: 615–622. diversification: effects on innovation and firmGeringer JM, Beamish P, daCosta RC. 1989. Diversifi- performance in product-diversified firms. Academy of cation strategy and internationalization: implications Management Journal 40: 767–798. for MNE performance. Strategic Management Journal Hoskisson RE, Hitt MA. 1990. Antecedents and perfor- 10(2): 109–119. mance outcomes of diversification: a review and cri-Geringer JM, Tallman S, Olsen DM. 2000. Product tique of theoretical perspectives. Journal of Manage- and international diversification among Japanese ment 16: 461–509. multinational firms. Strategic Management Journal Hoskisson RE, Hitt MA, Johnson RA, Moesel DD. 1993. 21(1): 51–80. Construct validity of an objective (entropy) categorialGhoshal S. 1987. Global strategy: an organizing measure of diversification strategy. Strategic Manage- framework. Strategic Management Journal 8(5): ment Journal 14(3): 215–235. 425–440. Hsiao C. 1986. Analysis of Panel Data. Cambridge Uni-Ghoshal S, Bartlett CA. 1990. The multinational corpo- versity Press: Cambridge, U.K. ration as an interorganizational network. Academy of Huber GP. 1991. Organizational learning: the contribut- Management Review 15: 603–625. ing processes and literatures. Organization ScienceGrant RM, Jammine AP, Thomas H. 1988. Diversity, 2(Special Issue): 88–115. diversification, and profitability among British man- Hymer SH. 1960. The International Operations of ufacturing companies, 1972–1984. Academy of Man- National Firms: A Study of Direct Foreign Investment. agement Journal 31: 771–801. MIT Press: Cambridge, MA.Gulati R. 1995. Does familiarity breed trust? The Jaccard J, Turrisi R, Choi KW. 1990. Interaction effects implications of repeated ties for contractual choice in multiple regression. In Quantitative Applications in in alliances. Academy of Management Journal 38: the Social Sciences, Vol. 72. Sage: Beverly Hills, CA. 85–112. Jacquemin AP, Berry CH. 1979. Entropy measure ofGupta AK, Govindarajan V. 1991. Knowledge flows diversification and corporate growth. Journal of Indus- and the structure of control within multinational trial Economics 27: 359–369. corporations. Academy of Management Review 16: 768–792. Jensen MC. 1986. Agency costs of free cash flow, cor-Haleblian J, Finkelstein S. 1999. The influence of porate finance, and takeovers. American Economic organizational acquisition experience on acquisition Review 76: 323–329. performance: a behavioral learning perspective. Johanson J, Vahlne JE. 1977. The internationalization Administrative Science Quarterly 44: 29–57. process of the firm: a model of knowledge devel-Hannan MT, Freeman J. 1977. The population ecology opment and increasing foreign market commitments. of organizations. American Journal of Sociology 82: Journal of International Business Studies 8: 23–32. 929–964. Kim WC. 1989. Developing a global diversification mea-Hannan MT, Freeman J. 1984. Structural inertia and sure. Management Science 35: 376–383. organizational change. American Sociological Review Kim WC, Hwang P, Burgers WP. 1993. Multinationals’ 49: 149–164. diversification and the risk-return trade-off. StrategicHastings DF. 1999. Lincoln Electric’s harsh lessons from Management Journal 14(4): 275–286. international expansion. Harvard Business Review Kobrin SJ. 1991. An empirical analysis of the deter- 77(3): 163–178. minants of global integration. Strategic ManagementHaveman HA. 1993a. Organizational size and change: Journal , Summer Special Issue (12): 17–37. diversification in the savings and loan industry after Kogut B, Kulatilaka N. 1994. Options thinking and plat- deregulation. Administrative Science Quarterly 38: form investment: investing in opportunity. California 20–50. Management Review 36: 52–72.Haveman HA. 1993b. Follow the leader: mimetic iso- Kogut B, Zander U. 1992. Knowledge of the firm, combi- morphism and entry into new markets. Administrative native capabilities, and the replication of technology. Science Quarterly 38: 593–627. Organization Science 3: 3838–3897.Hedlund G. 1986. The hypermodern MNC: a heterarchy? Kogut B, Zander U. 1993. Knowledge of the firm and Human Resource Management 25: 9–25. the evolutionary theory of the multinational corpo-Hedlund G. 1994. A model of knowledge management ration. Journal of International Business Studies 24: and the N-form corporation. Strategic Management 625–645. Journal , Summer Special Issue 15: 73–90. Lane C. 1995. Industry and Society in Europe: StabilityHennart J-F. 1982. A Theory of Multinational Enterprise. and Change in Britain, Germany, and France. Edward University of Michigan Press: Ann Arbor, MI. Elgar: Aldershot, U.K.Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)
  17. 17. Internationalization, Process Dependence and Profitability 653Lebas M, Weigenstein J. 1986. Management control: the Rugman AM. 1979. International Diversification and the roles of rules, markets and culture. Journal of Man- Multinational Enterprise. Lexington Books: Lexing- agement Studies 23: 259–272. ton, MA.Lewin K. 1936. Principles of Topological Psychology. Rugman AM. 1981. Inside the Multinationals: The Eco- McGraw-Hill: New York. nomics of International Markets. Croom Helm:Li J. 1995. Foreign entry and survival: effects of strate- London. gic choices on performance in international markets. Shleifer A, Vishny RW. 1991. Takeovers in the ‘60s and Strategic Management Journal 16(5): 333–351. the ‘80s: evidence and implications. Strategic Man-Malnight TM. 1995. Globalization of an ethnocentric agement Journal Winter Special Issue 12: 51–59. firm: an evolutionary perspective. Strategic Manage- Simon HA. 1959. Theories of decision-making in eco- ment Journal 16(2): 119–141. nomics and behavioral science. American EconomicMalnight TM. 1996. The transition from decentralized Review 49: 253–283. to network-based MNC structures: an evolutionary Stearns LB, Allan KD. 1996. Economic behavior in perspective. Journal of International Business Studies institutional environments: the corporate merger wave 27: 43–65. of the 1980s. American Sociological Review 61:Miller D. 1993. The architecture of simplicity. Academy 699–718. of Management Review 18: 116–138. Stopford JM, Wells LT. 1972. Managing the Multina-Murtha TP, Lenway SA, Bagozzi RP. 1998. Global tional Enterprise. Longman: London. mind-sets and cognitive shifts in a complex Sutcliffe KM. 1994. What executives notice: accurate multinational corporation. Strategic Management perceptions in top management teams. Academy of Journal 19(2): 97–114. Management Journal 37: 1360–1378.Nelson RR, Winter SE. 1982. An Evolutionary Theory of Tallman S, Li J. 1996. Effects of international diversity Economic Change. Harvard University Press: Cam- and product diversity on the performance of multi- bridge, MA. national firms. Academy of Management Journal 39:Nohria N, Ghoshal S. 1994. Differentiated fit and 179–196. shared values: alternatives for managing headquar- Tsai W. 2000. Social capital, strategic relatedness and ters–subsidiary relations. Strategic Management Jour- the formation of intraorganizational linkages. Strategic nal 15(4): 491–502. Management Journal 21(9): 925–940.Penrose E. 1959. The Theory of the Growth of the Firm. Vachani S. 1991. Distinguishing between related and Basil Blackwell: London. unrelated international geographic diversification:Perlmutter HV. 1969. The tortuous evolution of the multi- a comprehensive measure of global diversifica- national corporation. Columbia Journal of World Busi- tion. Journal of International Business Studies 22: ness, January/February: 9–18. 307–322.Phillips ME. 1994. Industry mindsets: exploring the cul- Venkatraman N, Ramanujam V. 1986. Measurement of tures of two macro-organizational settings. Organiza- business performance in strategy research: a compar- tion Science 5: 384–402. ison of approaches. Academy of Management ReviewPrahalad CK, Bettis RA. 1986. The dominant logic: 11: 801–814. a new linkage between diversity and performance. Vermeulen F, Barkema HG. 2001. Learning through Strategic Management Journal 7(6): 485–501. acquisitions. Academy of Management Journal 44:Reynolds PD. 1986. Organizational culture as related to 457–476. industry, position and performance. Journal of Man- Vernon R. 1966. International investment and interna- agement Studies 23: 333–345. tional trade in the product cycle. Quarterly JournalRonen S, Shenkar O. 1985. Clustering countries on atti- of Economics 41: 255–267. tudinal dimensions: a review and synthesis. Academy of Management Review 10: 435–454.Copyright  2002 John Wiley & Sons, Ltd. Strat. Mgmt. J., 23: 637–653 (2002)