07447556988

CHAPTER 2.

FOREIGN DIRECT INVESTMENT
AND THE MULTINATIONAL CORPORATION

2.1. INTRODUCTION
International business activity is by no means a recent phenomenon. The lives of
Phoenicians and Carthaginians, in the ancient world, were deeply dependent on international business. This economic activity included foreign direct investment (FDI), joint ventures and strategic alliances, among other forms of internationalisation (Moore and Lewis, 1999). Several multinational corporations (MNEs) can also be identified in
Europe in the middle ages and in the beginning of the modern era (Dunning, 1993a;
Jones, 1996).
The origins of modern international business activity however, are associated with the industrial revolution. Modern MNEs, in particular, have their roots in the massive international movement of factors that took place in the nineteenth century (Dunning,
1993a: p.99). Resource-seeking was the most common motivation of FDI in this period, even if by 1850 many firms had already crossed the Atlantic, in both directions, in what can be defined as market-seeking investment (Dunning, 1993a: p.100; Jones, 1996: p.5).

8

Despite the presence of FDI, most foreign investment in the nineteenth century - and indeed until the late 1940s – was portfolio capital. As a result, international business activity was largely ignored in economic theory until the late 1950s. On the one hand, the phenomenon did not have a major perceived economic impact. It was widely assumed that
MNEs were a passing post-war phenomenon originating in the United States (Jones, 1996:
p.3). On the other hand (and probably more importantly), the neo-classical theory, based upon perfect markets and the international immobility of factors, did not easily incorporate multinational activity.
The growth of FDI (and of the MNEs themselves) that followed World War II emphasised the inadequacy of the neo-classical theory to explain the phenomenon and the need for a whole new approach. The volume of FDI not only grew substantially, it started to reduce its concentration in primary goods, and to be increasingly directed towards the production of knowledge-based products in other developed countries (Buckley and
Casson, 1976: p.36). Furthermore, important changes in the organisation of international business were taking place, in particular, the development of horizontal MNEs and the new Japanese vertical foreign investments (Dunning, 1979: pp.270-2; 1993a, pp.126-7)1.
Despite its late arrival, international business literature (and in particular that on FDI) proliferated with increasing speed. The publication of the product cycle theory by
Raymond Vernon (1966) was followed by extensive research on the determinants of foreign production, in particular by scholars at the Harvard Business School led by Vernon himself. In the mean time, John Dunning brought a copy of Stephen Hymer’s 1960 PhD thesis to the University of Reading where, together with the work of Charles Kindleberger, it had a major impact. The two approaches of the ‘Reading School’ - the “internalisation theory” (Buckley, Casson, Rugman, Hennart) and the “eclectic paradigm” (Dunning) provided a consistent explanation of the reasons why firms choose to own production and trading facilities abroad. Furthermore, scholars at the University of Uppsala (Johanson,
Wiedersheim-Paul, Vahlne) started investigating the internationalisation process of individual firms, widening the scope of the new discipline.

Until World War II Japanese outward FDI was dominated by trading and financial companies (Dunning, 1993a:
p.124).
1

9

2.2. THE DETERMINANTS OF FOREIGN DIRECT INVESTMENT
Over a quarter of a century ago, Dunning (1973: p.289) observed that “[t]here are few branches of economic analysis which are not directly relevant to an understanding of the origin and growth of multinational enterprises”. The result was a wide range of approaches to international business, often dependent on the researchers’ backgrounds.
2.2.1. Capital theory
Until the 1950s, international direct investment was entirely explained within the traditional theory of international capital movements. Like other forms of international investment, FDI was seen as a response to differences in the rates of return on capital between countries. This suggestion was reinforced by the empirical observation that
American firms (the major source of FDI in the 50s) obtained a higher rate of return from their European investments than at home (Mundell, 1960). However, the differential rate of return hypothesis did not resist the inversion in that relationship registered in the 1960s, which was still accompanied by increases in US investment in Europe (Hufbauer, 1975).
Neither did it receive much empirical support2.
Hymer (1960) was the first to expose the deficiencies of this approach. He claimed that the differential rate of return hypothesis was not consistent with several observed characteristics of international investment. First, the United States combined net outflows of FDI with net inflows of portfolio capital. Second, flows of FDI in both directions between two countries were not rare. Third, many subsidiaries complemented the inflow of direct investment with capital borrowed in local markets. And, finally, manufacturing companies were at the time far more important in international direct investment than financial firms. Furthermore, an international difference in expected returns is not sufficient to induce FDI (Caves, 1982: p.25). Under perfect markets, an increase in the short run profits of firms in one country would not induce international investment.
Instead, it would attract new entrants that would eliminate any excess profits. Perfect markets and MNEs are not compatible (Hymer, 1960; Kindleberger, 1969; Hufbauer,
1975).

2

For a survey of empirical tests see, for example, Agarwal (1980: pp.741-2). See also Caves (1996: p.26).

10

Somewhat more refined than the differential rate of return hypothesis is the portfolio approach, developed in the 1960s using a Tobin/Markowitz stock adjustment model3. The portfolio approach assumes that part of the excess profits that should be earned in foreign markets are simply rents for higher risk associated with this alternative use of capital. As recently as 1992, Brainard and Tobin4 proposed a model in which FDI is simply one of the alternatives to portfolio investment. The rates of return of the different alternative investments are matched with an element of risk in the choice between (imperfectly) substitutable assets to build an efficient portfolio. However, the introduction of a risk correction element, more than being insufficient to eliminate the theoretical drawbacks of the underlying theory, highlights its deficiencies. In fact, Hymer’s criticisms of the differential rate of return hypothesis (see above) fully apply to the portfolio theory as well.
Furthermore, MNEs can provide a cheap international diversification of a portfolio, but only at a cost: the difference between the (rigid) international mix provided and each investor’s optimal mix. And this is very likely to off-set the initial cost-advantage. Finally, the portfolio hypothesis cannot explain the differences between industries’ propensities to invest abroad (Agarwal, 1980; Taveira, 1984).
According to Dunning (1973: p.299), the reason why portfolio theory can only partially explain direct foreign investment is that it ignores that “direct investment does not involve changes in ownership. It does, however, involve the transmission of factor inputs other than money capital, viz. entrepreneurship, technology, and management expertise, and is likely to be affected by the relative profitability of the use of these resources in different countries as that of money capital”. Furthermore, MNEs are not necessarily profits maximisers. Even if they are, there is no reason why they should forcibly seek higher profits on FDI than on domestic investment (Agarwal, 1980: p. 743).
Also in the capital theory tradition is the risk diversification hypothesis (Rugman, 1975,
1979; Lessard, 1976). The argument is that the international diversification of portfolios is a way of reducing the firm’s risk. This makes the MNE a vehicle for geographical diversification of investments. Caves (1996) explains, however, that although the empirical evidence shows that investors recognise the value of international diversification (p.160), the diversification of MNEs is more likely to result from investments that were propelled

Tobin (1958), Markowitz (1959). See Dunning (1973: pp. 300) and Agarwal (1980: p.745) for references to the application of portfolio theory to FDI.
4 Cited in Jong and Vos (1994a: p.9)
3

11

by other motives (p.21). Indeed, the geographical distribution of the portfolios of existing
MNEs, very much concentrated in highly correlated countries, is very different from what is suggested by the portfolio diversification hypothesis (Buckley, 1988: p.83).
2.2.2. The International Trade Tradition
It is certainly no surprise that International Trade economists were among the first to study the FDI phenomenon. Foreign production can be a substitute for exports, as it can influence the terms of trade and thus change the whole pattern of specialisation. However, in the neo-classical world of the Heckscher-Ohlin tradition there is no space for foreign direct investment. Any disequilibrium in the prices of goods or factors across countries brought about by different factor endowments would be immediately corrected by international movements of goods (the Samuelson theorem).
2.2.2.1. Mundell and the Heckscher-Ohlin model
Mundell (1957) used an extension of the basic model to show that trade and capital movements can be substitutes, namely, that the introduction of tariffs would induce a flow of FDI towards the country where tariffs are imposed5. That is, the same way that restrictions to international movements of factors can be substituted by trade (the original
H-O model), restrictions to trade can be replaced by international movements of factors, in particular capital given the intrinsic imperfect mobility of labour.
In a way, these hypotheses based on the Heckscher-Ohlin model are not very different from those based on capital movements. As Taveira (1984: p.10) points out, in both cases
“FDI was analysed as a re-equilibrium device within a generally perfectly competitive economy”, a major limitation of the explanatory potential of both approaches.
2.2.2.2. Kojima’s ‘Macroeconomic Approach’
Also in the neo-classical factor endowments tradition is Kojima’s ‘macroeconomic approach’ (Kojima, 1973, 1978, 1982). Kojima tried to explain the distinctive character of trade-oriented Japanese FDI, obeying the principle of comparative advantages, vis-à-vis US investment conducted in an oligopolistic market structure, anti-trade oriented and damaging to both home and host countries in the long run (Dunning, 1993a: p.90).

Corden (1974) showed that tariffs are not the only impediment to trade that originates FDI from a situation of different factor endowments.
5

12

The basic theorem is that “Direct Foreign Investment should originate in the investing country’s comparatively disadvantaged industry (or activity), which is potentially a comparatively advantaged industry in the host country” (Kojima, 1982: p.2). If this is the case, Kojima argues, (pro-trade oriented, or Japanese) FDI and international trade are complementary and lead to a dynamic reorganisation in the international division of trade and the associated gains for all countries involved.
The role of FDI can thus be seen as to exploit the home country’s comparative advantages in intermediate inputs that are embodied in products whose final stages of production give a comparative advantage to the host country (Dunning, 1993a). This is a most significant suggestion: some factor endowments generate comparative advantages that are better exploited abroad. That is, firms can build their competitive advantages upon the home country’s specific location-advantages, but best exploit these advantages, partially or totally, abroad, an idea also developed by Dunning (1981a) and to be discussed later.
The macroeconomic approach was the target of many criticisms. Its neo-classical perfect market assumptions are clearly a major limitation, for they ignore economies of scale, product differentiation and other forms of market failure (Dunning, 1993a; Jong and Vos,
1994b). It is not that Kojima is not aware of them. But being unable to distinguish firm level economies of scale from plant level economies (Buckley, 1983b: p.97), he fails to understand that in the presence of market failure hierarchies can improve the international allocation of resources (Dunning, 1993a: p.90).
Another limitation of the macroeconomic approach is its excessive concern with the distinction between the positive impact of Japanese “pro-trade oriented” FDI and the US
“anti-trade oriented” FDI. Kojima’s belief is that US FDI in technologically advanced industries was premature and doubly damaging. On the one hand, it did not fit the hostcountry’s factor endowments and associated comparative advantages. On the other hand, it prematurely eroded the United States’ technology-based competitive advantages.
Cantwell (1991), however, argues that export-oriented FDI is not necessarily better than import-substituting foreign investment. The latter can have highly positive spill-over effects. Its total long-term impact on trade can be positive. Furthermore, if of an enclave kind, export-oriented FDI will have little impact on the host-country’s technology and entrepreneurial levels.

13

Buckley (1983b, 1985, 1991) and Clegg (1987) further suggest that, because of its narrow assumptions, Kojima’s theory is not applicable even to most Japanese direct investment.
“Japanese-type” investment is not more frequent in Japan than in other developed countries (Buckley, 1983b: p.346). As Japanese MNEs matured, the distinction between
Japanese- and American- type FDI eroded. And Japanese import-substituting investments in Europe and in the US are certainly not less important than Japanese export-oriented
FDI in (mostly) Asian countries (Clegg, 1987; Cantwell, 1991). Buckley (1985) goes as far as to claim that even the analysis by Kojima and Ozawa (1984) of the Sogo-Shosha,
Japan’s traditional general trading companies, implicitly rejects the macroeconomic model.
2.2.2.3. The Product Cycle Model
Another stream of work that partially builds upon the factor-endowments tradition is the one that takes into account the role of innovation and the diffusion of knowledge. Posner
(1961), Hufbauer (1966), Vernon (1966), Hirsch (1967) and Wells (1972) are probably the most important references, with the product cycle theory, normally associated with
Vernon, being the model that better describes the role attributed to MNEs in the interaction between technology, international production and trade.
The argument is that technological development generates changes in the products’ factors intensity, thus changing the comparative advantages of countries. The role of demand, first discussed by Linder (1961), is also taken into account. Domestic demand can be an incentive to innovate, while international demand similarity facilitates exports. In a world with important technological and market barriers to trade (Hufbauer, 1966. Vernon, 1966),
MNEs are the most likely institutions to organise the production and distribution of goods with an international demand for which the most efficient production location is changing over time.
The Product Cycle described that American endowments of highly skilled labour and
R&D resources, matched with a highly sophisticated demand, prompted constant innovation among US firms. The consequent technological leadership was the basis for US exports and permitted the development of US multinationals which engaged in importsubstituting FDI in other developed countries. As products and technology matured, these advantages were progressively eroded, and US companies were forced to move to new products and technologies. These are then replaced by imitation-driven producers based, first, in other developed countries and, later, in developing countries. What was not clear

14

in the first versions of the product cycle (usually designated as Mark I) was whether the maturation process would drive out the production of US firms or simply production in the US.
Clegg (1987: p.24) claims that “[the product cycle] is not, in itself, a complete theory of
DFI as it does not explain the ownership of production”. Not least because the competitive advantage of firms is frequently associated with country-specific advantages
(Dunning, 1993a). Clegg (1987: p.26) adds that “the product cycle is primarily a theory of new FDI, and it has little to say on the extensions of existing investments by a mature foreign-investing nation”. Nevertheless, Dunning (1973: p.307) defends that “[trade based] models are of special interest in that they emphasise the role of innovations in forging new trade patterns within an imperfectly competitive environment, conditions which are the seed-bed of growth of the modern ME”. The trade approach has the merit of highlighting the fact that FDI is but one alternative to service foreign markets. Furthermore, it postulates “the distinctive character of the ME as an owner of resources in different countries compared with national firms”.
The Mark I Product Cycle received much empirical support from studies covering the
1950s and 1960s. But Vernon (1971: p.108) himself acknowledged that “by 1970, the product cycle model was beginning in some respects to be inadequate as a way of looking at the US-controlled multinational enterprises”. The successive revisions of the model Product Cycle Mark II (Vernon, 1974, 1979) - drove it very close to the HymerKindleberger approach (Buckley, 1981) - see section 2.3.2.
2.3. MARKET IMPERFECTIONS AND INDUSTRIAL ORGANISATION
The bases for a whole new approach of international production based on market imperfections were laid by Hymer (1960). However, his work was largely ignored until
Kindleberger (1969) published his own research6. Hymer’s work is clearly in the industrial organisation tradition - his major concern is with the organisation of production rather than trade flows - and largely inspired by Bain’s (1956) theory of barriers to entry in domestic industries7.

Hymer’s 1960 Doctoral dissertation was not to be published until 1976.
Cantwell (1991: p.22), however, observes that “in Hymer’s original version it was a theory of the firm and of the behaviour of the firm rather than a theory of industrial organization in the modern sense”.

6

7

15

2.3.1. The Hymer-Kindleberger hypothesis
The Hymer-Kindleberger hypothesis suggests that, because foreign firms have necessarily some disadvantages vis-à-vis domestic firms (e.g., knowledge of the market, communication), they must possess some firm-specific advantages if they are to engage in foreign production (Hymer, 1960; 1968). Furthermore, foreign direct investment is not about the transfer of capital - this could be supplied to local firms using other forms of international financing. It is about the international transfer of proprietary and intangible assets - technology, business techniques, and skilled personnel (Hymer, 1960: p.69). Hymer
(1960, 1968) claimed that the existence of FDI is exclusively due to the imperfection of the international markets for these assets. The firm “internalises or supersedes” these market failures through direct investment (Hymer, 1960: p.48).
The problem facing prospective international firms was summarised by Dunning (1973:
p.313): “there are two primary determinants of the amount of international production.
The first is the extent of the market in each country and the second is the competitiveness of foreign affiliates vis-à-vis indigenous and non-resident firms”. That is, the aim of any analysis should be “to identify both the location and ownership of firms”8.
A second key element in the Hymer-Kindleberger approach is why firms should choose to exploit their ownership advantages through direct investment rather than exporting, licensing, or other forms of international markets servicing. Buckley and Casson (1976:
p.68) and Rugman (1980: p.370), among others, claim this was never clearly explained by
Hymer. This was, nevertheless, implicit in Hymer’s (1960) original work, and extensively discussed in a later paper9. Hymer (1968: pp. 966-970) seems to believe that FDI is the most efficient internationalisation strategy, in particular when compared with licensing; if the advantage is based on technology or on some intangible asset, FDI was considered the most likely solution to maximise profits. Three reasons were presented: (i) the firm’s advantage may be very difficult to price; (ii) FDI eliminates the costs of defining and managing a licensing agreement; (iii) it is simply not possible to sell oligopolistic power.
Hymer (1960, 1968) viewed FDI very much as a way of defending and reinforcing market power in oligopolistic industries. In this, it is fully supported by Caves’s (1971) analysis of vertically integrated firms. Caves’s (1971: p.10) explanation of vertical FDI is an implicit
Italics in the original.
Hymer (1968). Apparently, the very existence this paper, published in French, was widely ignored until the middle
1980s.
8

9

16

assertion that multinationals can not only exploit perceived market imperfections, but they can use their ownership advantages to create market imperfections themselves. This is, nevertheless, an element only fully understood and integrated in the theory of international investment by Buckley and Casson (1976). However, Caves (1971: p.9) seemed to believe that imperfect markets for knowledge associated with product differentiation were the key for horizontal FDI (p.6). Despite being part of the truth, this is a restrictive view of the reasons behind horizontal foreign direct investment.
Before moving on to the next section, it is interesting to make a brief note on the surprising consistency between Hymer’s market power view of FDI and the mainstream
Marxist approach to foreign investment, or neo-imperialism10. The marxist argument is that the level of concentration (“monopolisation”) of the industries in capitalist countries generates very high profits. However, since oligopolistic collusion imposes restrictions on the re-investment of those profits at home, they must be invested abroad. Despite the difference in emphasis, this does not differ much from Hymer’s explanation of the role of oligopolies in the existence of FDI. Nevertheless, the Marxists tend to ignore the competitiveness of oligopolies that was central in Hymer’s approach. Instead, they emphasise the collusive anti-competition aspect of market power. As a result, the two approaches reach rather different conclusions: the neo-imperialists conclude that the expansion of MNEs (mostly from developed countries) into new (usually less developed) locations is nothing else but one more vector of the expansion of imperialism and yet another vehicle for the underdevelopment of the “Third World”.
2.3.2. The internalisation approach
Despite the invaluable contribution of Hymer, Kindleberger and Caves, the credit for transforming internalisation into a full paradigm of international production is usually attributed to Buckley and Casson (1976). These scholars did not simply complement previous work; they re-centred the analysis by building upon the theory of the firm (Coase,
1937). Looking at the firm as an alternative institution to markets, their theory “views the
MNE as a special case of the multiplant firm” (Buckley and Casson, 1976: p.36).
Buckley and Casson’s (1976) assertion that MNEs are typically both vertically and horizontally integrated led them to a model centred on the relationship between knowledge, market imperfections and the internalisation of markets for intermediate

17

goods. This comprehensive treatment of vertical and horizontal FDI is possible in so much as “the vertically integrated firm internalises a market for an intermediate product, just as the horizontal MNE internalises markets for proprietary assets” (Caves, 1996: p.13).
Additionally, internalisation will happen - and MNEs will grow - only as far as the benefits, including those associated with the barriers to new entrants, are not outweighed by the costs of communication, co-ordination and control, and the ‘foreignness’ inevitably associated with vertical and horizontal integrated firms. Rugman (1980, 1985) goes as far as to claim that this made internalisation a (the?) general theory of FDI, which will be discussed later.
The internalisation theory evolves from the concept of market failure. Some transactions are more efficiently performed inside the firm than in the market. Buckley and Casson
(1976: pp.37-38) specified five types of market imperfections that call for internalisation:
− when the co-ordination of resources over a long period is needed;
− when the efficient exploitation of market power requires discriminatory pricing;
− when bilateral monopoly produces unstable bargaining situations;
− when the buyer cannot price correctly the (usually intangible) goods on sale, or when public goods are involved;
− when government interventions in international markets create incentives for transferpricing.
Buckley and Casson (1976: p.39) listed several markets where internalisation is very likely to happen: perishable agricultural products, intermediate products in capital-intensive manufacturing processes, and raw-materials geographically concentrated11. However, these were secondary in the analysis. As with Hymer, at the centre of the analysis were the imperfections in the markets for knowledge12. These were ideal to illustrate why internalisation is the most efficient vehicle to exploit a proprietary advantage without putting at risk the monopoly it represents to the firm.
Also relevant in the internalisation theory is the perception that the firm is able to internalise externalities even when no market existed before: “the actions of firms can
See Jenkins (1987: p.27) for references.
Casson (1982: p.20) put it in different words: “MNEs will predominate in R&D-intensive industries, in resourcebased industries, and when the international division of labour is inhibited by fiscal intervention which can be avoided by transfer-pricing”.
10

11

18

replace the market or alternatively can augment it” (Buckley, 1981: p.9). That is, internalisation includes a theory of how new knowledge is created, a major departure from the Hymer-Kindleberger approach (Clegg, 1987: p.20).
The dichotomy replace/augment emanates from different connotations of ‘internalisation’ and has important welfare implications: “Internalization of a market refers to the replacement of an arm’s length contractual relationship by managerial co-ordination within the firm.
Internalization of an externality, however, refers to an improvement in social efficiency achieved by removing a defect or distortion in the price system” (Casson, 1987: p.36).
Several other authors made important contributions to the development of the internalisation theory. Horst (1971) presented the first microeconomic model of the choice between investing abroad and exporting from the home base. His model of the horizontally-integrated MNE demonstrated Hymer’s suggestion that FDI can be a strategy to enforce collusion. Horst (1972) also distinguished for the first time between ownership and location advantages, to be introduced in the next section (Clegg, 1987: p.32; Caves,
1996: p.54).
Aliber (1970, 1971) proposed a variant of the Hymer-Kindleberger model based on the existence of different currency areas. He argues that firms from countries with strong currencies can borrow at lower cost, which enables them to engage in risky investments in weak-currency areas. Aliber did not try to create a general theory of FDI. His model can be seen as the suggestion that firms internalise imperfections in the capital and exchange rate markets, as they do with any other market failure.
Johnson (1968, 1970) was the first to suggest that knowledge is a public good with nearzero social cost but non-zero private cost. This is the reason why the firm better exploits its knowledge-based advantages through internal markets, as fully addressed in Buckley and
Casson (1976). Magee (1977a, 1977b) extended Johnson’s work to build the notion of
“industry technology cycle”, largely inspired by the product cycle theory. He argued that the incentive for firms to internalise the market for technology varies over time. New technologies are more likely to be internalised (Magee used the expression
‘appropriability’), but as the technology matures licensing becomes increasingly attractive.
The licensing of a mature technology is easier to price and cheaper to monitor, thus
The emphasis was, nevertheless, different, since Hymer overlooked the concept of transaction costs and emphasised market failure (Dunning and Rugman, 1985: p.229; Casson, 1987: p.6).
12

19

reducing the risks and costs associated with the non-internalisation of the firm’s ownership advantage (Rugman, 1981).
Knickerbocker (1973) found that because of oligopolistic behaviour foreign subsidiaries tend to be clustered. Firms tend to follow competitors in their internationalisation decisions, a behaviour also found in the case of domestic diversification (Lamfalussy,
1961). This showed “that it was not just locational variables that determined the spatial distribution of the economic activity of firms but their strategic response to these variables and to the anticipated behaviour of their competitors” (Dunning, 1993a: p.72).
This notion of oligopolistic behaviour is at the centre of the revision of the product cycle, known as ‘mark II’. The emphasis of the “new” product cycle theory has moved away from technological development and international allocation of industries to strategic behaviour and how erected barriers to entry support international oligopolistic structures
(Vernon, 1974). In this new version, import-substituting FDI was expected as the product matured (as well as the technology, as Magee would put it). This strategic move intends to prevent damaging price wars. Because it is compelled by security rather than efficiency, the welfare outcome is not necessarily a world first-best.
The similarity with the internalisation theory discussed above is evident. However, and despite all the common ground, it must be stressed that a significant difference in approach still exists between product cycle mark II and internalisation. While the latter is a theory of the (international) firm, the former places itself at an industry level of analysis. It is a theory of international location based on oligopolistic behaviour, not a theory of the nature of the international firm.
Cantwell (1991: p.30) suggested an important distinction between the product cycle
“competitive international industry approach” and the “market power school” of the
Hymer-Kindleberger tradition. “While the market power school suppose that, in general, internationalization lowers the extent of competition and increases collusion amongst firms, competitive international industry approaches share the view that in general the growth of international production tends to be associated with rivalry and to sustain the process of technological competition amongst MNEs”. The observation seems to overlook, nevertheless, that in both cases FDI is both the response to and a vehicle of market change.

20

A rather different approach is Aharoni’s (1966) use of the behavioural theory of the firm to introduce the role of management and decision-making process in the explanations of the internationalisation of the firm. Recently, this approach has been substantially developed by Buckley (1993a, 1996a) who has been integrating the new developments in international business theory with those in the theory of strategic management.
2.3.3. The eclectic paradigm
According to Dunning (1979: p.274), the eclectic paradigm resulted from his dissatisfaction with existing theory of international production: the Hymer-Kindleberger approach, the product-cycle theory, and the internalisation theory. The three were considered to be partial explanations of international production. Henceforth, he proposed an alternative line of development which tried to integrate the existing theories in a general and ‘eclectic’ model in which “the subject to be explained is the extent and pattern of international production” (Dunning, 1991: p.124).
Dunning (1979: p.275) suggests that a firm engages in FDI if three conditions are satisfied:
(i) It possesses net ownership (O-) advantages vis-à-vis firms from other countries;
(ii) It is beneficial to internalise (I-advantages) those advantages rather than to use the market to pass them to foreign firms;
(iii) There are some location (L-) advantages in using the firm’s ownership advantage in a foreign location rather than at home.
The concept of ownership advantage is especially important to the eclectic paradigm, not least because it is probably what draws the line with the internalisation theory (Rugman,
1980, 1985; Casson, 1987). Dunning (1979: p.276) distinguished two sets of ownership advantages: those that result from an exclusive access to inputs, intangible assets or markets; and those directly associated with multinationality. Later (1983b), however, he put this distinction in slightly different, eventually more clear, terms. He distinguished between those ownership advantages that arise from the proprietary ownership of specific assets of the firm - asset (Oa) ownership advantages - which the firm can choose to internalise or not; and the ownership advantages that can only be exploited if internalised, since they result from the superiority of hierarchies vis-à-vis external markets in the common

21

governance of a network of assets located in different countries - transaction (Ot) ownership advantages13.
Dunning (1981a: pp.34-35) also considered necessary a systematic distinction between country (home and host), industry and firm determinants of the OLI characteristics: “the propensity of enterprises of a particular nationality to engage in foreign production will vary according to the economic at al. characteristics of their home countries and the country(ies) in which they propose to invest, the range and type of products they intend to produce, and their underlying management and organisational strategies”.
More recently, Dunning (1993a: p.79) added a fourth, firm-specific, condition to the basic three proposed in 1979:
(iv) Given the configuration of the OLI advantages facing a particular firm, the extent to which the firm believes that foreign production is consistent with its long-term management strategy.
One of the main criticisms of the eclectic paradigm is that it includes so many variables that it loses any operationality. Dunning (1991: p.125) partially accepts it, although he sees it as an inevitable consequence of trying to integrate the rather different motivations behind FDI in one general theory. He also accepts that the first versions of the OLI paradigm did not give full account of the dynamic interaction between the variables. The answer to this criticism was the Investment Development Cycle, or Path, which Dunning first introduced in 198114, and which will be discussed later.
2.3.4. OLI versus internalisation
Dunning (1993a: p.85) argues that “the [eclectic] paradigm is less an alternative theory of international production than one which pinpoints the essential and common characteristics of each of the mainstream explanations”. That is the reason why he renamed it ‘paradigm’ instead of the original ‘theory’. However, the claim that the eclectic paradigm has uniquely the global explanation of international production is not universally accepted. Rugman (1980), in particular, claims that internalisation is in itself a general

The actual expressions used in Dunning (1983b) were ‘asset-power’ and ‘transaction-power’ (p.334). The terms asset
(Oa) and transaction (Ot) ownership advantages only appear in subsequent works.
14 Dunning (1991: p.134, footnote 13; 1981a: p.30, footnote) refers that the notion of an Investment Development
Cycle was first proposed by him and Peter Buckley in 1975 at a conference of the UK Chapter of the Academy of
International Business, and again in 1978 with Peter Buckley and Robert Pearce in a similar conference.

13

22

theory of foreign direct investment15. He extensively analysed previous contributions to the theory of FDI to demonstrate that internalisation is the key element in all existing explanations. Hennart (1986) and Casson (1987) seem to agree.
Supporters of internalisation consider that the concept of ownership advantage is irrelevant in explaining multinational activity. Buckley (1983a) saw it as the consequence of applying static concepts to a dynamic issue - the growth of the firm. Casson (1987: p.135) went further, to argue that “Dunning’s eclectic theory implicitly denies the original powerful insight of Coase, which is that internalization is the raison d’être of the firm”.
Dunning’s distinction between asset and transaction ownership advantages may be seen as a concession to this criticism (Corley, 1992: p.11). But Casson (1987) admitted that the empirical work recognises the importance of ownership advantages.
Dunning’s interpretation is that the difference is one of semantics: “I accept that some ownership-specific advantages are the direct result of firms internalizing the market for their intermediate products across national borders. However, (…) I think it appropriate to refer to the benefit as an ownership-specific advantage and internalization as the modality by which this advantage is realized” (Dunning, 1991: p.132). Ownership advantages may be dynamic and volatile, but they are the factors that, by being internalised, allow firms to cross borders and become MNEs.
Dunning (1991) accepts that the internalisation theory has the leading explanation of why a firm should choose to engage in foreign investment. But he dismisses its capacity to explain the level, structure and location of all international production. Dunning’s claim is that for the internalisation theory to achieve that status all kinds of market imperfections would have to be considered in the approach, “in which case the theory loses much of its incisiveness” (p.120). Dunning sees the internalisation theory not as an alternative but as a very important contribution to his own approach. One, he admits, that considerably influenced the evolution of his own view of foreign direct investment (1991: pp.122-123).
The most important distinction between the two versions of the Reading School is probably the explicit reference to the role of macroeconomic variables in shaping international production. Since the very beginning, and despite the many subsequent developments, the internalisation approach is a theory of the firm that chose to cross national borders - a theory of the MNE. By contrast, the eclectic paradigm is a theory of
15

Although he seems to associate Dunning with the internalisation school (see Fina and Rugman, 1996: p.200)

23

FDI. It wraps the theory of the firm with the macroeconomic and socio-political environment in which the decisions are made: “The main difference between the determinants of intra-national and international production lies in the unique economic, political and cultural characteristics of separate sovereign states” (Dunning, 1993a: p.86).
2.3.5. Motives for foreign production
The motives for firms to engage in foreign production can be classified in four groups: natural resources seeking, market seeking, efficiency seeking and strategic asset seeking16.
Natural resources seeking FDI is justified by the fact that these resources – e.g. minerals, raw materials and agricultural products - tend to be location specific. The need to guarantee a cheap and safe supply of natural resources justified much of the FDI flows in the 1800s and early 1900s, largely from the most industrialised nations (i.e. Europe, USA and Japan) to the less developed areas of the globe (Dunning, 1993a: pp.110,124). Market seeking corresponds to FDI that aims at supplying the local market or markets in adjacent territories. It may represent a deeper involvement of the firm, following the success of exports, or the expansion of the firm to a wholly new market. Transportation costs and government regulations are the main reasons behind market seeking FDI. However,
Dunning (1993a: pp.58-59) suggested that strategic reasons may also be associated with this type of FDI. Some examples are to follow the firm’s clients in their foreign expansion, the need to adapt products to local conditions and tastes, or the reduction of transaction costs. Efficiency seeking FDI has two main forms. First, and probably the most frequent type, firms often seek to increase their cost efficiency by transferring production, totally or in part, to low labour costs locations. This is especially likely to happen in industries where unskilled or semi-skilled labour represents an important part of the production costs.
Common examples are US investment in Mexico’s maquiladoras, and investment in Portugal and Spain by north and central European countries (cf. chapter 5). The second type of efficiency seeking FDI corresponds to investment aimed at rationalising the operations of existing MNEs. The target may be the exploitation of comparative advantages in adjacent territories (e.g. following a process of economic integration, such as the creation of the
It should be noted that this classification differs from that of Dunning (1993a), which considered a category of
‘resource seekers’ that included the natural resources seekers, the search for cheap supplies of unskilled or semi-skilled labour, and the acquisition of technological capabilities, management or marketing expertise, and organisational skills
(Dunning, 1993a: p.57).
16

24

Single European Market, in 1992), or to exploit economies of scale and scope across borders. However, prior market seeking FDI or costs reducing FDI is a pre-condition for this variation of efficiency seeking foreign investment.
Finally, strategic asset seeking FDI is probably the fastest growing of the four motives for overseas investment (Dunning, 1994). Firms increasingly use FDI to obtain strategic assets
(whether tangible or intangible) that may be critical to their long-term strategy but are not available at home (see also section 2.3.3). In contrast to the other motives for FDI, strategic assets seeking investment does not imply the exploitation of an existing ownership advantage of the firm. Instead, FDI may be a vehicle for the firm to build the ownership advantages that will support its long-term expansion at home and abroad, as argued, for example, in the network literature (see section 2.4.1). Alternatively, strategic asset seeking investment may not involve strengthening the firm’s position, but rather to weaken the competitive position of its competitors (Dunning, 1993a: p.60).
2.4. A DYNAMIC APPROACH TO FOREIGN PRODUCTION
All the explanations of foreign production discussed so far are static approaches. Yet the choice of international production and management is essentially a dynamic issue.
According to Dunning and Rugman (1985: p.231), this intrinsic dynamism was already present in Hymer’s original work, in his treatment of ownership advantages. Hymer (1960) presented internationalisation as a way of enforcing market power, which implies an evolving world where there is no space for the notion of equilibrium. In a later work,
Hymer (1968) made even more explicit this dynamic bi-directional interaction between the internalisation of markets and market structure (Buckley, 1990).
However, the credit for the first consistent attempt to create a dynamic model of international production goes to Raymond Vernon (1966, 1974). The product cycle theory evolves around technological change and how it affects the distribution of production worldwide. The emphasis on dynamic interactions was reinforced in the revision of the model (Mark II) where oligopolistic behaviour, dynamic and in permanent disequilibrium by nature, takes the centre stage. Nevertheless, the classical tradition of general equilibrium that has always dominated economic thinking was not fully eliminated. The stages analysis suggests that the dynamic phases are periods of evolution between intermediate points of equilibrium. 25

Kojima (1982: p.8) expected to explain the dynamic effects of FDI with his macroeconomic approach. In a concession to his earlier critics, however, he admitted that the use of a comparative static method for a real dynamic model of international division of labour is restrictive. Nevertheless, Kojima and Ozawa (1985) insisted on the dynamic nature of their analysis of the impact of the international transfer of factors of production and goods in the welfare of countries. They believed that “a study of the creation and international dissemination of entrepreneurial endowments is the key to developing a theory of dynamic comparative advantage” (Kojima and Ozawa, 1985: p.136). But their method was still ‘comparative static’.
In his criticism of Kojima’s model, Buckley (1985, 1991) exploited the dynamic elements associated with internalisation. He argued that internalisation provides a greater cooperation between the different units of the firm, which in the long run stimulates R&D and is likely to provide dynamic welfare improvements (Buckley, 1985: p.119). Cantwell
(1989, 1991) has a very similar reasoning, only with the emphasis on the role of technology accumulation. Nonetheless, the internalisation theory lacks a truly dynamic approach.
Buckley (1990: p.663) seems to agree when he argues that there is a need to integrate approaches that pay attention to “the dynamics and disequilibrium at the levels of the firm, markets and international competitors” both in the strategic trade theory and in the theory of international business.
It must be said that attempts were made to incorporate dynamic elements in the theory of international production. Partially influenced by Aliber (1970), Buckley and Casson (1981) analysed the foreign market servicing decision of firms. In their model, firms switch between modes of foreign market servicing with different fixed and variable costs in response to changing market conditions. However, Buckley (1983a) considered that the assumptions required by the model made it too complex to be operational.
More importantly, Buckley (1983a) demonstrated that the incorporation of dynamic elements in the analysis of international production refuted the proposition, central to
Hymer’s original work, that local firms have an advantage over foreign entrants. Buckley
(1983a) argued that a stepwise analysis of foreign market entry highlights the importance of elements intrinsic to multinationality that “make the established MNE a radically different competitor than a first-time foreign entrant” (p.48). The determinants of foreign expansion for new foreign investors can differ from those of established MNEs.

26

The increasing importance of dynamic issues is the central element in Buckley and
Casson’s (1998) ‘state of the discipline’ discussion. They consider that uncertainty and market volatility made MNEs’ flexibility the leitmotif of the new research agenda. Foreign market entry decisions can no longer be seen as a sequence of one-off events, but as a continuos systemic process (Buckley and Casson, 1998: p.22). New issues must be brought to the fore: international joint ventures; cooperation and business networks; entrepreneurship and corporate culture; organisational change.
2.4.1. The Scandinavian School
With an intrinsic dynamic approach to international business, the Scandinavian school
(also called “Uppsala” or “internationalisation” - Johanson and Wiedersheim-Paul, 1975;
Johanson and Vahlne, 1977; Johanson and Mattsson, 1988; Vahlne and Nordstrom, 1988) was largely developed in the 1970s from the empirical observation of the internationalisation process of individual (mostly Swedish) firms. From their empirical observations, Johanson and Wiedersheim-Paul (1975) and Johanson and Vahlne (1977) concluded that firms gradually develop their international operations by a process of incremental knowledge and commitment.
As in Hymer (1960), it is believed that, because they have little or no knowledge of the local conditions, foreign firms are at a disadvantage vis-à-vis local competitors. Although
‘objective’ knowledge about foreign countries can be bought by the company,
‘experiential’ knowledge can only be obtained through direct experience (Johanson and
Vahlne, 1977). Only the effective presence in foreign countries provides this critical element if the firm is to become an efficient player in international markets. Moreover, the process of foreign expansion is influenced by the firm’s past experience, the size of potential markets and, most importantly, the firms’ psychic distance to each potential host country. The latter is defined by factors such as the differences between home and host countries in terms of language, culture, political systems, level of education, and level of industrial development. Because of the correlation between cultural and geographic distance, psychic distance is also normally strongly associated with geographic distance17.
Typically, internationalisation starts with exports via independent representatives (agents), followed by the establishment of sales subsidiaries and, eventually, productive subsidiaries.
Johanson and Vahlne (1977: p.33) defined psychic distance as "the sum of factors preventing the flow of information from and to the market".
17

27

This is clearly a process of increasing resources’ commitment, as well as progressive knowledge acquisition. The fact that foreign subsidiaries are frequently established through the acquisition of former agents or by contracting key persons in the agents’ structure
(Johanson and Vahlne, 1977: p.33) is consistent with the knowledge acquiring view.
At the same time, the knowledge acquired in neighbouring countries (in terms of psychic distance), where internationalisation is likely to start, will permit the progressive expansion to countries increasingly further apart. That is, economies of scope in the learning process allow the firm to expand to new foreign countries, ever more distinct from the home country. These economies of scope also permit the firm to overcome the restrictions imposed by limited managerial capacities, which would not permit the firm to enter several foreign markets simultaneously (Casson, 1994). Furthermore, the impact of the internationalisation process in the firm’s organisational capacity, human resources and organisational structure (see Welch and Luostarinen, 1988) will probably enable it to jump stages after certain critical knowledge of international markets is obtained. This is particularly evident when the firm expands its operations to countries psychologically far from the home country, but close to others where it is already established.
On the face of this, internationalisation is no more than “the consequence of a process of incremental adjustments to changing conditions of the firm and its environment”
(Johanson and Vahlne, 1977: p.35)18. However, a fully dynamic approach to international production was not provided by the Scandinavian researchers until the mid 1980s. The original model only tried to explain early stages of internationalisation, ignoring competitive factors that change over time, in particular international competition.
Johanson and Mattsson (1988) and Vahlne and Nordstrom (1988) argued that in order to analyse situations where both the firm and the market are highly internationalised it is necessary to look at industrial markets as networks of relationships between firms19.
Knowledge and resource commitment remained the cornerstones of the network approach. However, it considers that the internationalisation of the firm depends on its capacity to build long term links with other firms in foreign networks (Johanson and
Mattsson, 1988: p.296). This network of relationships permits the creation of a capital of trust that reduces transaction costs and increases cooperation in the development of new
18

Aharoni (1966) is recurrently cited by the authors of the internationalisation school.

28

products and technology. In other words, it represents a specific competitive advantage even when it is an unintended by-product of the firm’s short term options (Vahlne and
Nordstrom, 1988: p.262). But the network is also in permanent change, and the firm’s position in it requires constant investment.
Vahlne and Nordstrom (1988: p.262) argued that a successful entry in an international network depends on the firm possessing some specific (e.g. technological) advantage (an
Oa-advantage in Dunning’s parlance). But once established in the network, the knowledge of the market and the special relationship with suppliers and customers becomes an advantage in itself, allowing the firm to maintain its international position even if the original advantage erodes20.
According to this approach, internationalisation depends on the firm’s network relationships rather than on firm-specific advantages (Coviello and McAuley, 1999: p.227).
Henceforth, firms may not internationalise to exploit existing ownership advantages (cf. section 2.3.3). Instead, internationalisation may be the vehicle to access foreign strategic assets that will permit to offset prior deficiencies in the firm’s ownership advantages (cf. section 2.3.5 on strategic asset seeking FDI). The network is, in this sense, a facilitating element. Both Fujita (1995) and Gomes-Casseres (1997) found evidence that smaller firms
(less likely to possess strong ownership advantages) rely on network linkages to build up their ownership advantages and to gain economies of scale and scope (Chen and Chen,
1998: p.446).
2.4.2. Modes of foreign market servicing
The internationalisation school presented foreign market entry as an incremental process.
However, the choice was limited to that between a subsidiary (FDI) and a contractual arrangement (licensee or agent). The latter was expected to precede the former (Vahlne and Nordstrom, 1988: p.258), as well as purely commercial FDI being expected to precede productive FDI. Furthermore, in a process of incremental involvement, joint ventures represent an intermediary stage between contractual arrangements and wholly owned international projects.

Arguably, the original analysis applies essentially to investors from small countries (Johanson and Mattsson, 1988:
p.299). Firms from countries with big domestic markets may be large enough to start internationalisation with a big productive FDI project.
20 The convergence with Buckley’s (1983a) suggestion that multinationality is an advantage in itself (Dunning’s Otadvantages) is obvious.
19

29

Buckley and Casson (1976) used a costs-benefits analysis to suggest a very similar international involvement path. Their claim was that, in normal conditions, the fixed costs associated with licensing are lower than those resulting from FDI. They are, however, higher than exports because of the need to guarantee that the licensing agreements are respected by the licensees. Since the opposite happens with variable costs, market servicing tends to follow the sequence: exporting - licensing - FDI. Buckley and Casson
(1981) added that the switch in modes of market servicing is also affected by the life-cycle of the product, the firm’s familiarity with the foreign market, and the firm’s degree of internationalisation. Rugman (1981) also examined the choice between exporting, FDI and licensing. He was, however, very much concerned with the appropriability problem (Magee (1977a, 1977b) and believed that licensing is a risky modality. “The very existence of the MNE is threatened by premature or otherwise inappropriate licensing” (Rugman, 1981: p.70).
Hence, he concluded that licensing will only take place in highly mature industries, which results in that the sequence between the three foreign market servicing strategies will be exporting–FDI-licensing. This negative view of licensing, in particular in the earlier stages of the product cycle, is shared by Vahlne and Nordstrom (1988: pp.258-259).
In fact, it seems that Rugman provided a very detailed analysis of a special case of Buckley and Casson’s (1981) model - when one of the modes (licensing) is inefficient21.
Alternatively, it may be suggested that Buckley and Casson (1976) underestimated the fixed costs associated with licensing. Nevertheless, Rugman (1981: p.74) concedes that the growth of standardised products and a better government regulation are increasing the use of non-equity forms of international involvement (licensing, joint ventures). In other words, better regulation reduces the costs of licensing, increasing its attractiveness at any stage of maturity.
2.4.3. The Investment Development Path
The Investment Development Cycle, or Path, was introduced as a dynamic approach to the OLI paradigm (Dunning, 1981a: p.34). However, contrarily to the eclectic paradigm where the macroeconomic variables are simply one level of analysis, the investment development path is largely a macroeconomic approach (Cantwell, 1991: p.39).
This possibility is generically discussed by Buckley and Casson (1981: p.80). They also suggested that the choice between foreign market servicing alternatives is affected by the product cycle (p.85).

21

30

The investment development path suggests an association between a country’s level of development (proxied by GDP per capita) and its international investment position (net outward FDI stock per capita). The basic hypothesis is that, as the country develops, the conditions facing domestic and foreign companies change. This will have an impact on the flows of inward and outward FDI. However, inward and outward FDI affect the economic structure as well. In other words, there is a dynamic interaction between the two. The IDP also accepts that governments can influence the country’s conditions and, consequently, FDI flows and domestic firms’ ownership advantages, a notion new to the mainstream theory of FDI.
According to the IDP, countries evolve through five stages of development (Dunning,
1981a, 1981b, 1986b; Tolentino, 1987; Dunning and Narula, 1996b):
Stage 1
The first stage is associated with pre-industrialisation. Countries in this stage will not attract any foreign investment, except probably for a few companies eventually interested in exploiting existing natural resources, but with little or no integration in the national economy. Very small domestic markets, inadequate infrastructure, a poorly educated labour force and undeveloped commercial and legal frameworks are some of the factors that explain this low attractiveness. On the other hand, domestic companies do not possess any significant ownership advantages, and outward FDI will be nil. Dunning
(1981a: p.38) suggests that, if they exist at all, O-advantages are probably best exploited through other forms of international contracting (e.g., minority direct investment, portfolio resource flows or exports).
Governments at this stage usually have two sets of actions. They try to improve basic infrastructure and to upgrade human capital; and they adopt macroeconomic policies that are intended to change the structure of domestic markets and industries - import protection and export promotion are two examples.
Stage 2
The combination of national policies pursued by the government will eventually create some location specific advantages. In consequence, inward FDI starts to rise, probably attracted by an emergent domestic market in consumer goods, but also in transport, communications and construction (including public demand in infrastructure). Frequently,

31

this happens in response to tariffs imposed by the government. As in stage 1, exportoriented FDI will probably exist in natural resources-based industries. Some vertical integration into labour-intensive activities upstream in the value-chain can also be expected if and when basic infrastructure has been provided. Labour-intensive manufacturing and tourism are other sectors likely to attract foreign investment at this stage.
Outward direct investment at this stage will be low, reflecting the scarce ownership advantages of domestic firms. It is likely that existing O-advantages have been developed in industries connected with natural resources or other primary activities that managed to produce semi-skilled and moderately knowledge-intensive consumer goods. Hence, despite its initial low level outward FDI will start to rise as domestic firms engage in marketseeking FDI in (probably less developed) adjacent territories and, more important to the development of their O-advantages, in strategic asset-seeking investment in developed countries. The domestic government frequently has an active role in inducing these early internationalisation attempts. Dunning (1993a) argues that the combination of domestic and foreign investment that results from the country’s improved L-advantages will generate agglomerative economies and increase labour productivity. This will positively affect both domestic firms O-advantages (and decrease foreign firms’) and the country’s
L-advantages themselves. Hence, it is claimed, “in these initial stages of development, the role of government is especially important” (Dunning, 1993a: p.88)22.
An important characteristic of this stage is that the combination of fast growing inward
FDI with only exploratory outward foreign investment will make the country an increasingly net receiver of FDI. Or, in other words, the country’s net stock of foreign investment is increasingly negative.
Stage 3
The development of domestic firms’ O-advantages and increasing production costs associated with higher wages will translate, sooner or later, into a reduction in the rate of growth of inward FDI and an increase in the rate of growth of outward direct investment.
As a result, net inward investment per capita will start to fall. That is, comparative advantages in labour-intensive industries will deteriorate, creating an incentive to search for new less developed locations for these industries. At the same time, stronger O-advantages

32

of local firms make them more able to cope with an increasingly exigent domestic demand prompted by the rising incomes, and with foreign competition. Larger markets also mean more opportunities for economies of scale, favouring the development of domestic firms’
O-advantages. Market seeking outward FDI to both less and more developed countries is also expected, as it is strategic assets-seeking investment in stage 4 and 5 countries.
Furthermore, the changing O-advantages of domestic firms will be decreasingly associated with the home country’s specific characteristics or government policies and more with the possession of intangible assets and knowledge by the firms themselves. In other words, Oadvantages, at first largely country-specific, will become progressively firm-specific. FDI induced O-advantages, or the advantages resulting from managing and co-ordinating geographically dispersed assets (Ot-advantages), make their appearance. O-advantages also become an active element in the reshaping of the country’s L-advantages, side by side with government policies and economic growth. These L-advantages will now be defined by a large domestic market, a growing stock of human capital, and a stronger technology capacity. In response, import-substituting inward FDI will be progressively replaced by efficiency-seeking production.
Dunning (1981a: p.41) still recognises a significant role for governments at this stage.
Governments have the tasks of further reducing market imperfections and of encouraging a deeper integration between domestic and foreign firms. Dunning claims that governments’ policies should have two distinct aims. On the one hand, to attract foreign investment to industries where domestic firms are unable of exploiting existing Ladvantages. On the other hand, to provide incentives for the internationalisation of domestic firms in those industries where they already possess significant O-advantages and the country’s L-advantages are weak or eroding. This may mark the beginning of the country’s international investment specialisation. In any case, “structural adjustment will be required if the country is to move to the next stage of development” (Dunning and
Narula, 1996b: p.6).
Stage 4
Countries in stage 4 are those that became net outward investors, with outward FDI still growing faster than inward FDI. It means that domestic firms now possess the ownershipadvantages to compete in any domestic or foreign market. They grew in size and they
22

Dunning (1993a: p.88) cites also Porter (1990) and Ozawa (1989).

33

diversified both geographically and in terms of industries. At this stage, Ot-advantages, those resulting from multinationality, are far more important than Oa-advantages.
Production processes are capital and knowledge intensive. The reasons to engage in outward FDI will also diversify. In labour-intensive industries, domestic firms will continue to engage in efficiency-seeking FDI in (less developed) countries with lower wages.
Outward investment to overcome trade barriers will be found in countries in any stage of development. Rationalised and strategic asset-seeking investment in other countries in stages 4 and in countries in stage 5 will take place in innovatory industries.
Following a tendency felt since the very first stages of the IDP, the country’s L-advantages are now almost entirely based on created assets. Consequently, inward FDI will include market- and asset-seeking direct investment from countries in lower stages of development, but it will mostly be rationalised and strategic asset-seeking investment from other stage 4 countries. Intra-industry production is a consequence of the growing similarity in the O-advantages of firms from countries at this stage, and it generally follows prior growth in intra-industry trade. In part, it translates the increasing propensity by
MNEs to internalise trade and production (Dunning and Narula: 1996b: p.7).
At this stage, the role of government has changed. It still concentrates on improving market efficiency and reducing transaction costs. However, and more importantly, governments begin to take a more strategic intervention in supporting infant industries and reducing the economic and social impact of eliminating declining industries.
Stage 5
The existence of a fifth stage of the IDP to include the leading developed countries was first suggested by Dunning only in 1986 (Dunning, 1986b: pp.30-31). It resulted from the difficulty to explain the convergence and balancing of FDI stocks in most developed countries since the middle 1980s. Despite permanent high stocks of both inward and outward FDI, the net outward investment (NOI) position of stage 5 countries will revolve around zero, alternating between positive and negative balances according to the short term evolution of exchange rates and economic cycles.
Dunning (1986b)23 suggests that this is the combined influence of economic and technological convergence among the leading developed countries with the tendencies already described in stage 4: countries’ L-advantages are increasingly associated with

34

created assets, and firms’ O-advantages are more transactions-based and less assets-based.
As a result, cross-border trade and investment are essentially conducted inside the MNEs.
International investment flows of any country in stage 5 will be more dependent on the strategies of its MNEs and of MNEs from other countries in stages 4 and 5 than on differences between the countries themselves (which are very few anyway). These MNEs will increase their commitment to rationalising their international production networks
(which they certainly started in previous stages) with investment in other developed (stages
4 and 5) countries. They will also continue to direct FDI to less developed countries, mostly in natural resources and labour intensive industries. At the same time, stage 5 countries will be the recipients of market-seeking and strategic asset-seeking investment from countries in lower stages of development.
It should be noticed that it is implicit in the description of stage 5 that no single country has an advantage over the other developed economies. MNEs, alone and independently of the domestic or host country’s location advantages, are the dominant force in shaping international production and trade. MNEs increasingly behave like “mini-markets”
(Dunning and Narula, 1996b: p.8) and to some of them the whole concept of home country is becoming meaningless, as they transform themselves from Multinational
Enterprises into Transnational Corporations.
Governments, nevertheless, retain a role in the dynamic economic restructuring. Buckley
(1996b: p. 2) suggests that “a fundamental role of government is to seek to appropriate some of the rents earned by transnational firms”. As firms assume a greater importance in shaping the world economy, national governments are increasingly assuming the role of strategic oligopolists. They must take into account the behaviour of MNEs, but also that of other governments (Dunning and Narula, 1996b). Inevitably, some governments will be more able to play the dual oligopolistic game than others.
2.5. CONCLUSION
It is clear from the above review that different approaches to international business resulted in a wide ranging body of literature. The reasons why individual firms engage in international activities are the centre concern. And the key seems to be market imperfections, which prompt companies to internalise cross border activities. In many
23

See also Dunning and Narula (1996b: pp.7-9).

35

respects, this is not very different from what firms do inside national borders. As such, internationalisation can be seen as just another dimension of the growth of the firm
(Buckley, 1993b). Nonetheless, this is a dimension unlike the others. On the one hand, countries differ in their legal, political and cultural characteristics, which generates a whole set of managerial problems. On the other hand, multinationality changes the very nature of the firm, and can be in itself a source of competitive (ownership) advantages24.
The issue of change, or the intrinsic dynamic nature of the internationalisation process, was of particular appeal to the Scandinavian school. As any disadvantage faced by any firm, those associated with doing business in a foreign country are neither permanent nor universal. Internationalisation is a learning process. It can be managed through a process of progressive commitment of resources, starting in more familiar countries and moving to ever more distant ones. If the internalisation of ownership advantages explains why
MNEs exist at all (Dunning, 1991), only this dynamic approach to internationalisation can describe the process of the development of a multinational corporation.
However, in order to understand the way international production is organised worldwide a last critical element is needed. And that is the importance of locational factors. Only the interaction between the internalisation of ownership advantages with location advantages can explain the configuration of MNEs’ activity worldwide. This seems to be a description of the eclectic paradigm (Dunning, 1979), except that it cannot be dissociated from the dynamic elements of the investment development path (Dunning, 1981a, 1981b, 1986b).
The aim of this research project is to analyse the competitiveness of Portugal from the perspective of the country’s position in the international production network. From what was described above, this will entail the investigation of the locational advantages associated with foreign and domestic firms operating in Portugal, as well as the process of internationalisation of domestic firms. The eclectic theory and in particular the investment development path will constitute an appropriate framework of analysis.

That the geographic distribution of the firm’s activities can be in itself an ownership advantage was recently restated by Dunning (1998a).
24

Cited: in Jong and Vos (1994a: p.9) 3 MNEs, very much concentrated in highly correlated countries, is very different from what is suggested by the portfolio diversification hypothesis (Buckley, 1988: p.83). 2.2.2.1. Mundell and the Heckscher-Ohlin model Mundell (1957) used an extension of the basic model to show that trade and capital approach’ (Kojima, 1973, 1978, 1982). Kojima tried to explain the distinctive character of trade-oriented Japanese FDI, obeying the principle of comparative advantages, vis-à-vis US investment conducted in an oligopolistic market structure, anti-trade oriented and damaging to both home and host countries in the long run (Dunning, 1993a: p.90). Corden (1974) showed that tariffs are not the only impediment to trade that originates FDI from a situation of different factor endowments. comparatively advantaged industry in the host country” (Kojima, 1982: p.2). If this is the case, Kojima argues, (pro-trade oriented, or Japanese) FDI and international trade are a comparative advantage to the host country (Dunning, 1993a). This is a most significant suggestion: some factor endowments generate comparative advantages that are better abroad, an idea also developed by Dunning (1981a) and to be discussed later. product differentiation and other forms of market failure (Dunning, 1993a; Jong and Vos, 1994b) level economies of scale from plant level economies (Buckley, 1983b: p.97), he fails to understand that in the presence of market failure hierarchies can improve the international allocation of resources (Dunning, 1993a: p.90). Cantwell (1991), however, argues that export-oriented FDI is not necessarily better than import-substituting foreign investment 13 Buckley (1983b, 1985, 1991) and Clegg (1987) further suggest that, because of its narrow “Japanese-type” investment is not more frequent in Japan than in other developed countries (Buckley, 1983b: p.346) FDI in (mostly) Asian countries (Clegg, 1987; Cantwell, 1991). Buckley (1985) goes as far as to claim that even the analysis by Kojima and Ozawa (1984) of the Sogo-Shosha, one that takes into account the role of innovation and the diffusion of knowledge. Posner (1961), Hufbauer (1966), Vernon (1966), Hirsch (1967) and Wells (1972) are probably the discussed by Linder (1961), is also taken into account. Domestic demand can be an incentive to innovate, while international demand similarity facilitates exports with important technological and market barriers to trade (Hufbauer, 1966. Vernon, 1966), MNEs are the most likely institutions to organise the production and distribution of goods