This article has three objectives:
(1) to explore reasons for the continuance of agents/distributors and sales subsidiaries as international market entry and development modes and channels of distribution in the UK machine tool industry;
(2) to facilitate an understanding of mode evolution in the sector; and
(3) to interpret the empirical results within the frameworks of internalization/transaction cost analysis, internationalization and establishment chain models, and network theory.
Data were collected in Spring and Summer 1993 by means of in-depth interviews with 14 agents/distributors and 14 foreign owned sales subsidiaries, and interpreted qualitatively.
Market entry modes and channels of distribution - literature review
There is a continuing debate on the determinants of the market entry modes and channels of distribution utilized by companies as they internationalize. Theoretical approaches basically divide into three, namely internalization/transaction cost analysis, internationalization and establishment chain models, and interaction network theory. Much of the interest of international business economists has been the multinational enterprise (MNE) and the conditions under which (mainly) production activities are established and evolve overseas; the internalization/transaction cost approaches (Buckley and Casson, 1976) and the eclectic paradigm (Dunning, 1988) are the principal conceptual frameworks. Buckley et al. (1990) have widened the application of internalization theory to the non-production functions of stocktaking, distribution, generating customers and transport; in this way the work is brought closer to that of channel researchers who have applied transaction cost analysis to explain the choice between the internalization of marketing and distribution functions and the employment of outside agents and intermediaries. For Anderson and Coughlan (1987) integration may be preferred when the firm possesses specialized knowledge and when agents are difficult to find. More generally, the authors argue that integration of the channel function is likely when:
(1) products require a high service level, are differentiated, less mature and are closely related to the company's core business;
(2) the level of transaction-specific assets in the salesforce is high; and,
(3) psychic distance is low.
Integration may be encouraged by the transaction difficulties identified by Williamson (1975), namely, those arising from opportunism, bounded rationality, and asset specificity, with environmental uncertainty as an additional element. However, trust can reduce transaction costs by restraining opportunism: by reducing complexity, trust should also alleviate some of the difficulties associated with bounded rationality (Andaleeb, 1992). When firms are unable to trust their agents, hierarchical forms of transactions are the most likely outcome.
An alternative framework, again focusing on the emergence of the MNE, postulates an internationalization process which develops in four stages according to an establishment chain: no regular export activities; exports via agents; exports via sales subsidiary; and production via foreign subsidiary (Johanson and Valhne, 1977; Johanson and Wiedersheim-Paul, 1975). According to this behaviouralist model, the firm would make additional incremental resource commitments as it gained experience from current activities; the stages represent possible indicators of the operation of the process model. Johanson and Valhne (1990, p. 12) consider that "the internationalization process, once it has started, will tend to proceed regardless of whether strategic decisions in that direction are made or not". Comparing the internalization/transaction cost and internationalization model, there is a fundamental difference between the assumption of rational decision making in the former and the behavioural responses to uncertainty avoidance in the latter. The critical factors in the internationalization model are lack of knowledge of the foreign market and an absence of established relationships to customers and other parties. It is argued that unless these constraints are included in an assessment of internalization costs and benefits, the internalization/transaction cost paradigm cannot explain the shifts in mode which are demonstrated by the internationalization model. Even if transaction costs are high in the early stages of internationalization, as they might be for technologically complex products, the company might be unwilling to move beyond agency arrangements because of lack of knowledge of the market and network relationships. In general there is substantial empirical evidence both in support of or consistent with the internationalization model and, increasingly, against.
A further major stream of work relevant to this article has developed from international industrial marketing, focusing on interactions, relationships and networks. The interaction approach (Hakansson, 1982) draws on a range of existing streams of thought including inter-organizational theory and institutional economics. It applies concepts such as risk reduction, power and dependence, distribution channel behaviour and industrial buying behaviour. The approach is fundamentally descriptive, pragmatic and flexible in its analysis. Thus authors such as Turnbull (1987) have been critical of the mechanistic sequential model of internationalization, observing moves both "forwards" and "backwards" in international marketing organizational forms. An interesting criticism from the 1986 research of Turnbull and Valla (1986) was that while investment in a sales subsidiary may produce certain competitive advantages, it could also create barriers or constraints which limit flexibility of action in disordered, volatile industrial markets.
The interaction approach may be encompassed within the broader concept of industrial networks, with co-operation, bonding and complementarity of objectives being stressed. Reductions in uncertainty and increased stability could result, as well as, on the other side, increased dependence. Johanson and Valhne (1990) have attempted to extend the internationalization model by linking it to the concept of industrial networks. The authors accept the evidence that the internationalization model is less appropriate in turbulent, high technology industries (Lindqvist, 1988). Their explanation is that entrepreneurs in such firms have international networks of colleagues who influence internationalization behaviour and facilitate the rapid establishment of subsidiaries. More recently (Johanson and Valhne, 1992) these same authors have argued that many firms enter new foreign markets almost blindly, with market entry emerging out of the interplay between actors in the foreign market and the local firm, including social exchange processes.
The concepts of power, conflict/co-operation and trust (derived from resource dependency theory) are powerful ones and have been extensively discussed in the channel literature (Ennew et al., 1993; Gaski, 1984) and included within different theoretical frameworks. As noted earlier, trust has been regarded as having the potential to reduce transaction costs: for example, might trusting relationships between principal and agent lead to the maintenance of market transactions and reduce the need for internalization? In other instances the outcomes would be more uncertain, however, as in the application of coercive or exercised power. The industrial network's approach also highlights the importance of social relationships, and co-operation as well as conflict and dependence. If power is asymmetrically distributed and/or partners believe they are not receiving an equitable share of the benefits from a relationship, then network breakdown is likely. As this …