The model extends the insights of internalization theory, and draws on concepts from the economics of industrial studies of foreign direct Li investment (FDI)have become much more ambitious in scope over the last 30 years. In the 1960s, the main focus of the Hymer-Kindlebergertheory (Hymer 1976, Kindleberger, 1969) and the prodEImpirical *Peter organization. A special feature of the model is the distinction between investment in production facilities and in distribution investment facilities – an important practical distinction that has been overlooked in much of the international business literature.
The strength of competition from indigenous rivals is emphasized as a determinant of entry strategy into both production and distribution. uct cycle theory (Vernon 1966) was exporting versus FDI. In the 1970s the internalization approach identified licensing, franchising and subcontracting as other strategic options. The resurgence of mergers and acquisitions and Director of the Centre of J. Buckley is Professor of International Business International Business Studies at the University of Leeds. **Mark C. Casson is Professor of Economics at the University of Reading.
A preliminary version of this paper was presented to the Annual Conference of the Academy of International Business (U. K. chapter) at Leeds University in April 1997, the Conference on International Firms, Strategic Behaviour and International Location at the University of Paris I, Pantheon-Sorbonne in May 1997, and the Joint Annual Conference of the ESRC Industrial Economics Network and the International Economics Study Group at Nottingham University, June 1997. The authors are grateful to the discussants for their comments.
The paper has also benefited from the suggestions of three anonymous referees. JOURNAL OF INTERNATIONAL BUSINESS STUDIES, 29, 3 (THIRD QUARTER 1998): 539-562. 539 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKET ENTRY in the 1980s – often as a “quick fix” route to globalization – highlighted the choice between greenfield ventures and acquisitions. At the same time, the growing participation of U. S. firms in international joint ventures (IJVs)drew attention to the role of co-operative arrangements.
In the 1990s, the role of FDI in “transitional” or “emerging”economies (East and Central Europe, China, Vietnam, etc. ) has broughtback into focus some of the classic issues of the 1960s: The “costs of doing business abroad,” and the importance of “psychic distance. ” It has renewed interest in the general questions as to why some modes of entry offer lower costs than others, and why certain circumstances seem to favor certain modes over others. Linking all these issues together generates a high degree of complexity.
Although the eclectic theory has been regularly revised and updated to accommodate the changing foci of applied research, it is too much of a “paradigm” or “framework” and too little of a ”model” to provide detailed advice on research design and hypothesis testing (Dunning, 1980).
Complexity appears to have created a degree of confusion amongst scholars, which only a formal modelling exercise can dispel. The model presented below has three distinctive features. First, it is based on a detailed schematic analysis that encompasses all the major market entry strategies.
In existing literature, most strategies are appraised as alternatives to exporting, or as alternatives to greenfield FDI. It is unusual to see a direct comparison between, say, licensing and joint ventures, or between franchising and subcontracting. The present model permits any strategy to be compared with any other strategy. It is therefore 540 particularly useful when the leading strategies in contention do not include either exporting or conventional FDI. The second featureof the model is that it distinguishes clearly between production and distribution.
Historically, a large proportion of initial FDI relates to foreign warehousing and distribution facilities. Production facilities only come later, if at all. The distinction is obvious in empirical work, but it has not been properlyreflected in theory up until now. The result has been some confusion as to how theory should be applied to situations in which investment in distributionhas a prominentrole. Finally, the model takes account of the strategic interaction between the foreign entrant and its leading host-country rival after entry has taken place.
Following recent developments in industrial organizationtheory (as summarized, for example, in Tirole, 1988), it is assumed that the entrant can foresee the reaction of its rival, and take this into account at the time of entry. It is argued that this theoretical refinement is of the utmost practical importance in explaining the choice between greenfield investment and acquisition as entry modes. The model concentrates on FDI for market access reasons, and excludes resource-orientated FDI and offshore production.
OF HISTORICALDEVELOPMENT THE THEORY Much of the early literature on foreign market entry concerned the choice between exporting and FDI (for previous overviews, see Root, 1987; Young, et al. 1989; Buckley and Ghauri,1993).
The cost-based view of this decision suggested that the firm must possess a “compensating advantage” in order to overcome the “costs of foreignness” JOURNAL OF INTERNATIONALBUSINESS STUDIES This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions & C. PETER BUCKLEY MARK CASSON J. Hymer, 1976; Kindleberger, 1969).
This led to the identification of technological and marketing skills as the key elements in successful foreign entry (Hirsh, 1976; Horst, 1972).
This tradition of firm-specific advantages (Caves, 1971; Rugman, 1981) connects with the literature on core competences arising from the Penrosian tradition (Penrose, 1959; Prahalad and Hamel, 1990).
Sequential modes of internationalization were introduced by Vernon’s “Product Cycle Hypothesis” (1966), in which firms go through an exporting phase before switching first to marketseeking FDI, and then to cost-orientated FDI.
Technology and marketing factors combine to explain standardization, which drives location decisions. commitment to each market. Increasing commitment is particularlyimportant in the thinking of the Uppsala School (Johanson and Wiedersheim-Paul, 1975; Johanson and Vahlne, 1977).
Closely associated with stages models is the notion of “psychic distance,” which attempts to conceptualise and, to some degree, measure the cultural distance between countries and markets (Hallen and Wiedersheim-Paul, 1979).
For a more recent view see Casson, 1994. Non-Production Activities
In explaining foreign market servicing policies, the role of non-production activities must be made explicit. The location of research activities is widely debated, especially in relation to spatial agglomeration (Kogut and Zander, 1993).
There is also an extensive literature on the entry aspects of marketing and distribution (Davidson and McFetridge, 1980), much of it in a transactions cost framework (Anderson and Coughlan, 1987; Anderson and Gatignon, 1986; Hill, Hwang and Kim, 1990); Kim and Hwang, 1992; and Agarwal and Ramaswani, 1992).
Internalization
Buckley and Casson (1976) envisaged the firm as an internalized bundle of resources which can be allocated between product groups, and between national markets. Their focus on market-based versus firm-based solutions highlighted the strategic significance of licensing in market entry. Entry involves two interdependent decisions on location and mode of control. Exporting is domestically located and administratively controlled, foreign licensing is foreign located and contractually controlled, and FDI is foreign located and administratively controlled.
This model was formalised by Buckley and Casson (1981), and empirically tested by Buckley and Pearce Contractor (1984) and others. (1979), Mergers and Acquisitions Versus Greenfield Ventures Stopford and Wells (1972) examined takeovers versus acquisitions as part of the their analysis of the organisation of the multinational firm. The predominance of entry via takeovers in most advanced economies has stimulated a number of good empirical studies (Dubin, 1975; Wilson, 1980; Zejan, 1990; Hennart and Park, 1993), which have drawn on both the internalization perspective and the strategy literature (Yip, 1982).
Particular attention has been paid to the costs of adaptation and cultural integration that are encountered in 541 Stages Models of Entry The Scandinavian “stages”models of entry suggest a sequential pattern of entry into successive foreign markets, coupled with a progressive deepening of VOL. 29, No. 3, THIRD QUARTER,1998 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKETENTRY the case of mergers. The theoretical issues have recently been surveyed by Svensson (1996) and Meyer (1997).
Joint Ventures Versus Wholly Owned Subsidiaries The recent literature on IJVs is immense, and has spawned some innovative developments in internationalbusiness theory and much insightful empirical work based on extensive data sets (Contractorand Lorange,1988; Beamish and Killing, 1997).
Buckley and Casson (1988, 1996) summarize the conditions conducive to IJVsas: (i) the possession of complementary assets; (ii) opportunities for collusion, and (iii) barriers to full integration- economic, financial, legal or political (see also Beamish,1985; Beamish and Banks, 1987; Kogut, 1988; 1990).
Hennart,1988; and Contractor, The IJVliteraturehas focused particularly on partner selection, management strategy and the measurement of performance. Partnerselection is examined by Beamish (1987), who relates selection to performance, Harrigan (1988b), who examines partner asymmetries, and Geringer(1991).
Kogut and Singh (1987, 1988) relate partner selection to entry method. Managementstrategy in IJVsis analysed by Killing (1983) and Harrigan (1988), whilst Gomes-Casseres (1991) relates strategyto ownership preferences.
The performanceof IJVsis the subject of much debate. It cannot be assumed that joint venture termination indicates failure an IJV may end precisely analyses of IJV performance include Geringerand Hebert (1991), Inkpen and Birkenshaw (1994) and Woodcock, Beamish and Makino (1994).
Nitsch, Beamish and Makino (1996) relate entry mode to performance, and Gulati (1995) examines the role of repeated ties between partners as contributing to success – an interesting attempt to encompass “cultural”variables. CulturalFactors
The relationship between (national) culture and entry strategy is explicitly examined (using a reductionist version of Hofstede’s (1980) cultural classification) by Kogut and Singh (1988) (see also Shane, 1994).
Cultural barriers are utilized in an examination of foreign market entry by Bakema, Bell and Pennings (1996), and a “cultural learning process” is invoked by Benito and Gripsrud (1994) to help explain the expansion of FDI. Market Structure and Entry Strategy It is one of the contributions of this paper to introduce market structure issues into the modelling of entry decisions.
The relationship between entry behaviour and market structure was emphasized in Knickerbocker’s (1973) study of oligopolistic reaction, which set up a crude game-theoretic structure for competitive entry into key national markets. Flowers (1976) and Graham (1978) emphasized “exchange of threats” in their respective studies of European and Canadian investment in the United States, and two-way investment between the United States and Europe. Yu and Ito (1988) more recently examined oligopolistic reaction and FDI in the U. S. tyre and textiles industry. Graham (1992) laments the lack of JOURNAL OF INTERNATIONAL BUSINESS STUDIES ecause it has achieved its objectives. Similarly, the restructuringof joint ventures and alliances may indicate the exploitation of the flexibility of the organizational form, rather than a response to under-performance – see Franko (1971), Gomes-Casseres (1987), Kogut (1988, 1989), and Blodgett (1992).
Other 542 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions PETERJ. BUCKLEY MARK C. CASSON & attention to competitive structure in the international business literature, where the entrant is effectively a monopolist (Buckley and Casson, 1981).
Indeed, Casson’s (1985) study of cartelization versus multinationalization is one of the few economic models of multinational industrial organization available The Entrant 1. A firm based in a home country is seeking to sell for the first time in a foreign market. The emphasis on first-time entry makes it important to distinguish between the one-off set-up costs of an entry mode, and the recurrent costs of subsequent operation in that mode. It is assumed, unless otherwise stated, that recurrent operations take place in a stable environment. 2.
Foreign market demand for the product is infinitely elastic at a price P1, up to a certain volume at which it becomes totally inelastic. For example, each customer may desire just one unit of the product, which they value at Pl, and when everyone has bought that unit no more can be sold however far the price is dropped. The volume at which demand becomes inelastic is determined by the size of the foreign market,x. 3. The focus of the model on market entry makes it appropriateto distinguish between production activity (P) and distribution activity (D).
Distribution links production to final demand. It comprises warehousing, transport, and possibly retailing too. Distribution must be carried out entirely in the foreign market, but production may be located at either home or abroad. 4. The entrant’s production draws upon proprietary technology generated by research and development activity (R).
Effective distribution depends upon marketing activity (M).
Marketing involves investigating customers’ needs, and maintaining the reputation of the product by giving customers the service they require. 5.
The entrant has no foreign activity M at the time of entry, and consequently lacks market knowledge. This knowledge can be acquired through experi543 Summary Location costs, internalization factors, financial variables, cultural factors, such as trust and psychic distance, market structure and competitive strategy, adaptation costs (to the local environment), and the cost of doing business abroad are all identified in the literature as playing a role in determining firms’ foreign market entry decisions. The model which follows includes all these variables, and analyzes their interactions in a systematic way.
THE MODEL The model applies the economic theory of FDI presented in Buckley and Casson (1976, 1981), Buckley (1983), Casson (1991) and Buckley and Casson (1996) to the set of issues identified in the literature review above. Although the model involves a number of apparently restrictive assumptions, these assumptions can, if necessary, be relaxed, at the cost of introducing additional complications into the analysis. The assumptions are not so much restrictions upon the relevance of the model as indicators of key contextual issues on which every researcher into foreign market entry must pass udgement before their analysis begins. If some of the assumptions seem unfamiliar then it is because few researchers have actually made their assumptions sufficiently explicit in the past. VOL. 29, No. 3, THIRDQUARTER, 1998 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKETENTRY ence (learning from mistakes) at the time of entry, incurring a once-and-forall cost of entry, m. The knowledge can be obtained in other ways as well, as described below.
One of the keys to successful entry strategy is to acquire M in the most appropriateway. 6. The flow of technology from R to P defines the first of three “intermediate products” in the model. The second is the flow of marketing expertise from M to D. The third is the physical flow of wholesale product from the factory or production unit P to the distribution facility D. (The internal flow of information between R and M is not discussed as it is a fixed cost, which is the same for every form of market entry considered in the model. ) 7. Production at home means that the product must be exported.
Exporting incurs transport costs and tariffs that foreign production avoids. On the other hand, foreign production incurs additional costs of communicating the technology, e. g. , training foreign workers. Foreign production may also result in the loss of economies of scale. Exporting increases the utilization of the domestic plant, and allows it to be extended at low marginal cost. All of these factors are summarized in the net additional cost of home production, z, which is equal to transport costs and tariffs less savings on account of training costs and economies of scale. 8.
The firm may enter the foreign market either by owning and controlling * Pand D; * * P only; D only; or * Neither P nor D. In the second case, it uses an independent distribution facility, which is franchised to handle the product. In the third case, it either exports from its home pro544 duction facility, or subcontracts to an independent local facility. In the final case, the firm licenses an independent local firm to both produce and distribute the product. Because there is only one host-countryrival (see 14 below), the possibility that the firm could subcontractto one firm and franchiseanotheris ignored. . The transactioncost of operatingan external market is normally greaterthan that of an internal one. The availability of alternative incentive structures in an internal marketreduces the costs of haggling and default (Hennart, 1982).
Indeed, it is assumed in the present model that the transactioncost of obtaining marketingexpertise from an external consultant, rather than from the firm’s own M activity, is prohibitive. The entrant can tap into an established M activity only by franchising the local rival, forming a joint venture with the rival, or acquiringits distributionfacility. 0. The cost of external transfer of technology is also high, but acceptably so. One of the main problems in transferring technology is to monitor the output of the production process to make sure that the contractis being complied with. This is easier to do under a subcontracting agreement, where the product is “bought back,” than under a licensing agreement,where it is not. The transactions costs of a subcontractingagreement exceed the internal costs of technology transferby t1, whilst the costs of licensing exceed internal costs by t2-t1. 1. When the ownership of P differs from that of D then the flow of intermediate products between them is effected through an external market. When compared to the alternative of vertical integration of P and D, this incurs additional transaction costs, t3. 12. Entry of any type can be effected by either greenfield investment or BUSINESS STUDIES OF JOURNAL INTERNATIONAL This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions PETERJ. BUCKLEY MARK C. CASSON & cquisition. Under greenfield investment the firm uses its funds to pay for the construction of a new facility. Under acquisition it uses its funds to purchase the facility second-hand as a going concern instead. This is done by acquiring the equity in the firm which previously owned the facility. 13. An effective internal market requires a high degree of trust within the organization. This trust is not available immediately after an acquisition. It costs q1 to build trust in technology transfer when a P facility is newly acquired.
It costs q2 to build trust in the transferof marketing expertise when a D facility is newly acquired, and q3 to build trust in the transfer of intermediate product when either P or D (but not both) is newly acquired. The Host CountryRival 14. The firm faces a single local rival who previously monopolized the foreign market. At the time of entry, this rival operates as a fully integrated firm. It has the expertise, conferred by an activity M, which the entrant lacks. However, the local rival has higher costs because of inferior technology, on account of having no activity R. 15.
It is assumed that in all bargaining (for example, over an acquisition) the local rival plays an essentially passive role. The rival does not bargain for a share of the entrant’s profits, but simply ensures that it receives full opportunity earnings for the resources it surrenders to the entrant firm. The rival realizes that the entrant has a superior technology, and believes that when confronted with such a competitor its best strategy is to exit the industry by selling to the entrant those resources it wishes to buy, and redeploying the others to their best alternative use.
VOL. 29, No. 3, THIRDQUARTER, 1998 16. If the entrant uses the rival’s production facility, then a cost of adaptation is incurred. This is because the entrant uses a different technology from the rival, and equipment must be modified accordingly. This applies regardless of whether the entrant acquires the facility outright, or merely licenses, or subcontracts to the rival firm. However, the rival may have local production expertise, which the entrant lacks, providing savings to offset against the adaptation cost. The net cost of adaptation may therefore be negative.
A negative adaptation cost, in this context, signifies that the cost of adapting the entrant’s technology to local conditions using a greenfield plant is higher than the cost of adapting an existing local plant to the entrant’s technology. 17. By contrast, use of a rival’s D facility incurs no adaptation cost. This is because warehouses are normally more versatile than production plants. Use of the rival’s D facility always brings with it the marketing expertise associated with M. 18. The rival’s P and D facilities are the only existing facilities that can meet the needs of the market.
Other local firms cannot enter the market, and the rival firm itself cannot invest in additional facilities. Under these conditions, acquisition of either a P or D facility gives the entrant monopoly power: Acquisition of a D facility gives the entrant a monopoly of final sales, whilst acquisition of a P facility gives the entrant a monopoly of supplies to D. Greenfield investment, however, confers no monopoly power because it eliminates no rival facility: greenfield investment in D creates duopoly in the sourcing of final demand, whilst greenfield investment in P creates duopoly in the sourcing of D. 45 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKETENTRY 19. When the rival retains ownership of both its P and D facilities, then it remains a potential competitor. Although it may have switched some of its facilities out of the industry, it can, in principle, re-enter by switching them back again. If it has contracted out its P facility under a subcontracting arrangement, or contracted out its D facility under a franchising arrangement,then it can, in principle, re-enter competition when the agreements expire.
Under a subcontracting arrangement,the entrant and the rival remain potential competitors in the final product market, since each has its own distribution facility. Any attempt by the entrant to charge the full monopoly price would encouragethe rival to switch to producing its own output instead. The entrant must persuade the rival not to compete by reducing its price to a “limit price”P2 r4 lo, r4 r4 10 ‘. 4 C) ;-4 P4 cli cli C C r-i r-i ri N r r-4 r-4 r-i r-i Cli r-4 4-j u tj 40. co &4 0 0 -4 cl) P. 4 to a) *j U 0 . ,-I C” 4. 4-i a) cn :z 4C-0, 0 4C. a) cn -C cfj P4 co .b co m4-;4 C) 4-j . 4-i to 4-i C) tj C/3 0 110. “o 7z _4 44-i co to 0 fn 00 64 fn -4 0 1-4 7. ‘ 0 CO >1 4. ; 0 P–4 4. 0 40 fn 14;A U t co t 41), >1 jCd >1 40. fn 0 x 0 x tz 0 . P.. ( 0 C) 0 0 Izi la) 0 0 0 0 4. j j 0 –, 4-i CIJ . S -, 4u, a) o 0 o 64 u Cid ‘m r. ( cyi rn . ,-I –b cfj rn C) tj 0 0 . r. ( 0 cyi 0 r. ( 0 64 P4 cid 11 0 0 00 r4 -4 0 0 (2) z cid cn 0 0 O Ez, LF; 44 la) r4 cli cli 04 (6 o6 ci 6 r14 r4 r14 cli V-4 548 JOURNAL OF INTERNATIONAL BusiNESS STUDIES This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions
PETERJ. BUCKLEY MARK C. CASSON & TABLE 2 COMPARED WITH THE PROFITNORM STRATEGIES COSTS OF ALTERNATIVE C1. = cl. 2 = cl. 3 = rql + rq2 rq2 + ra + rq3 +S +rm cl. 4 = C2. 1 = rql rql +tt3 tj t1 z z z +rq3 t3 +ra +s +rm c2. 2= c3. 1 c3. 2= C4. 1z= C4. 2 = C5 = +t3 +rq2 + rq2 +t3 + rq3 + t3 +ra +ra +ra +s +rm +rm + s C6 = C7 = t2 rii + rj2 + ra + ra C8 = c9 = z rji tl C12. 1 = C12. 2 = rql rj1 rji +rj2 + rj2 + rj2 +rj2 +rj2 + rj3 + rj3 +rj3 +rj3 +rj3 + rj3 + rj3 + ra + ra +s/2 +s/2 +ra + ra + ra + s/2 + rm hand side of Table 1, and summarized schematically in Figure 1.
Six of these strategies have different variants generated by the fifth issue. These variants are indicated on the right hand side of the table. The figure distinguishes linkages involving the flow of information from R to P and M to D, and linkages involving the flow of physical product from P to D, and from D to final demand. Location is distinguished by the columns, and ownership by the rows. Ownership by the entrant is also identified by shading; facilities owned by the local rival are shown as clear. The strategies associated with each particular linkage are indicated by the numbers 1-12 in the figure.
Deriving the Profit Equations A profit equation for each variant of each entry strategy can be derived by applying the assumptions given above to the schematic illustrations in Figure 1. Certainelements of cost and revenue are common to all the profit equations, and it simplifies matters to net these out. This generates a set of summary profit equations in which profitability is expressed in terms of deviations from a profit norm. An appropriatenorm is the profit generated by pursuing strategy 1 under ideal conditions, in which the firm is already acquainted with the local market, and there is no indigenous rival.
The profit norm is the revenue generated by sales at the monopoly price Pl1 less 549 VOL. 29, No. 3, THIRDQUARTER,1998 This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKETENTRY the cost of greenfield foreign production, less the cost of greenfield foreign distribution, less the cost of internal technology transfer to a greenfield foreign plant, less the cost of internal transferof goods from production to distribution.
If the actual profits of each strategyare compared with this norm, then every strategy incurs some additional cost. The relevant cost expressions are given in Table 2. The subscripts applied to the cost symbol c refer to the strategies and their variants listed in Table 1. The variables on the right hand side have already been explained when introducing the assumptions of the model. Set-up costs are multiplied by the rate of interest to convert a once-and-for-all cost into a continuous equivalent. To see how the profit equations are derived, consider strategy 2.
This involves FDI in production, with sales being handled by the rival firm. There are two variants of this strategy,depending upon whether the production plant is acquired or not. The only international transfer of resources under this strategy involves technology, which moves across the column boundary from R to P. The transfer is internalized because no change of ownership is involved. Change of ownership only occurs where the flow of intermediateoutput from P to D crosses the row boundary. From D the product is distributed to the entire foreign market,as indicated by the flow fanning out from D.
The advantages of this particular strategy are two-fold. It internalizes the transfer of technology within the entrant firm, and it internalizes the transfer of marketing expertise within the local firm. This can only be achieved, however, by externalizing the flow of intermediate output, which generates the transaction cost premium 550 term t3, which appears in the expressions for both c2. 1 and c22. This is, in fact, the only term that is common to both expressions. The remaining terms are all accounted for by the difference between greenfield and acquisition methods of FDI.
The greenfield strategy avoids the cost a of adapting an existing plant to the needs of a new technology. Thus the term ra, which appears in the expression for c22, does not appear in the expression for c21. The greenfield strategy also means that the internal transfer of technology is not bedevilled by a lack of trust, which arises when the production facility is acquired instead. The cost of building trust in internal technology transfer, rq1, therefore appears in c22, but not in c21. The compensating advantage of the acquisition strategy is that it does not add to overall capacity in the foreign country.
Indeed, because the entrant faces a single local rival, acquisition of the rival’s production facility effectively prevents the rival from entering into competition with the entrant firm. Given that under strategy 2 the local firm retains control of distribution, it can threaten to source distribution from its own production plant instead of from the entrant’s plant. Although the entrant may be able to constrain this threat in the short term by signing an exclusive franchise contract with its local rival, in the long run this contract will expire, and the threat will reappear.
Only acquisition of one of the rival’s facilities can eliminate this threat altogether. This means that the greenfield strategy incurs a loss of revenue s compared to the acquisition strategy. Dominance Relations Theory predicts that the strategywith the lowest cost will be chosen. Which JOURNAL OF INTERNATIONAL BUSINESS STUDIES This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions PETERJ. BUCKLEY MARK C. CASSON & strategy is chosen depends on the relative magnitude of the different variables on the right hand side of Table 2.
The easiest way to understand the general properties of the solution is first to eliminate any strategies that are clearly dominated by others, and then to compare the remaining ones in terms of the major trade-offs involved. Whether strategies are dominated or not depends upon what restrictions are imposed upon the right-hand-side variables. So far, the only restrictions implied by the assumptions are m, r, s, ji, qi ti > 0 (i = 1,2,3) and t2-t1. In particular, the variables a and z are unrestricted in sign. Under these conditions, only two of the strategies are dominated, namely the bottom two in the table: C12. 1>C8; C12. >C8 C3. 1>c1. This means that the strategy of investing only in a greenfield distribution facility is inefficient compared to the strategy of investing in a greenfield production facility as well. Put simply, subcontracting production is not a good idea when the net cost of adapting existing plant to the new technology is positive. So far, no use has been made of restrictions on transactions costs. Suppose now that external market costs exceed the costs of building trust in internal markets after acquisition. In the context of production, this means that t1>rql from whence it follows that: c3. 2>c2. 2 c9>c11. The first inequality shows that subcontracting production in conjunction with the acquisition of a distribution facility is more costly than franchising distribution in conjunction with the acquisition of a production facility. The second inequality shows that subcontracting production in conjunction with a jointly owned distribution facility is more costly than acquiring a production facility in conjunction with a jointly owned distribution facility. These results underline the fact that high transaction costs in technology markets, combined with easy trust-building postacquisition, discourage subcontracting and favor acquisition instead.
The process of elimination through dominance can be continued by postulating that the cost of building trust is lower after an acquisition than it is within a joint venture: qi 0, then all the export strategies are dominated by equivalent strategies involving greenfield foreign production: c4. 1>c. 1l; c4. 2>C1. 3;C5>c21: c10>c11. 1 This illustratesthe importantpoint that location effects are independent of internalization effects in models of this kind. If the net cost of technological adaptation of existing production facilities is positive, a > 0, then it follows that: VOL. 29, No. 3, THIRDQUARTER,1998
This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions FOREIGN MARKETENTRY c7>c12; C11 1>Cl3 ; C11. 2>C13 It is inefficient to combine an IJVdistributionfacility with a production facility that is either wholly or jointly owned. Obviously, if the cost of building trust were thought to be lower in a IJVthen the inequalities would be the other way round, and the three acquisitions-based strategieswould be eliminated instead. It is not only inequality restrictions that can be used to generate dominance relations: equality restrictions can be used as well.
For example, if the costs of building trust after acquisition are the same in each internal market, qi=q (i=1,2,3), then: C1. 4>C1. 2 >C1M3 so that it is cheaper to combine greenfield production with greenfield distribution rather than with an independent distribution facility. The second restriction asserts that the transaction cost of the external intermediate product market exceeds the cost of building trust in that market following an acquisition; t3>rq3 . It follows that (given that ql=q2 from an earlier restric- tion): C2. 2>C13, o that it is cheaper to combine greenfield production with acquired distribution than to acquire production and franchise distribution instead. Properties of the Solution By carrying the process of elimination so far, only three of the original strategies are left in contention: 1. 3. greenfield production combined with acquired distribution; 2. 1. greenfield production combined with franchised distribution; and 6. licensing. The choice between these strategies is governed by six of the original variables: a, q, r, s, t2, t3. The solution is to choose: 1. 3. if q? (t3+s)12r,(t2Ir)+a 2. 1. if t3+s t3.
It follows that: C1. 1>C2. 1, 552 It can be seen that strategy 1. 3 is preferred wherever the cost of acquisition q is low. This is reasonable because 1. 3 is the only one of the three strategies that involves acquisition. Strategy 2. 1 is preferred when the transaction costs of the external market in intermediate output, t3, are low, and when the loss of monopoly profits from competitive distribution, s, is small. This is reasonable because strategy 2. 1 is the only one to involve an arm’s length sale of intermediate output, and the only one to leave JOURNAL OF INTERNATIONALBUSINESS STUDIES
This content downloaded from 115. 157. 199. 248 on Tue, 21 May 2013 09:59:52 AM All use subject to JSTOR Terms and Conditions PETERJ. BUCKLEY MARK C. CASSON & TABLE 3 COMPARATIVE STATICANALYSISOF THE EFFECTSOF CHANGES ON IN THE VALUESOF THE EXPLANATORY VARIABLES THE CHOICE BETWEEN THE THREEDOMINANTSTRATEGIES a q – s t2 t3 r 1. 3 2. 1 Acquisition Franchising + + + – + + + – + + Notes: a Adaptation cost of production plant. q Cost of building trust to access marketingexpertise through a newly-acquired distributionfacility. s Value of profit-sharingcollusion. 2 Additional transactioncost incurredby licensing technology. t3 Additional transactioncost incurred in using an external marketfor the wholesale product. r Rate of interest. the local rival in a position to compete. Strategy 6 is preferred when the transactions costs of licensing a technology, t2, and adapting local production facilities, a, are low. This is reasonable because the licensing strategy is the only one of the three to utilize existing production facilities; the other two use only existing distribution facilities instead.