1
A Contingency Model of Export Entry Mode Performance
Ian F. Wilkinson and Van NguyenSchool of Marketing, University of New South Wales Australia Sydney, NSW,Australia, 2052
Phone: 61-2-9385 3298
Fax: 61-2-9663 1985
Email:
April 2002
A Contingency Model of Export Entry Mode Performance
Abstract
A contingency theory of export mode performance is developed based on transaction cost analysis as well as production cost theory and the temporal nature of exchange relations. This is tested using the results of a nationwide survey of Australian manufacturers export entry modes. The Heckman two stage regression procedure is used to correct for self selection bias in the samples of firms using different export entry modes. The results demonstrate the importance of including factors affecting production costs as well as transaction cost in and that failing to correct for self selection bias can lead to misleading results and incorrect normative recommendations.
A Contingency Model of Export Entry Mode Performance
Introduction
The main theory used to explain the use of different market entry modes in foreign markets is transaction cost analysis (TCA). This theory is primarily normative yet most studies do not explicitly examine its performance effects. Instead, it is assumed that in competitive markets only the most profitable organisational forms will survive. After reviewing previous studies of TCA, including those focusing on foreign market entry modes, Rindfleisch and Heide (1997) note that this deficiency is a frequently expressed concern in the literature and that “The limited research on TCA’s performance implications makes it difficult to assess fully its theoretical value and empirical validity” (p47). They call for additional research which examines the performance implications of matching governance modes with structural conditions.
The purpose of the research reported here is to empirically examine the performance of export entry modes under different firm and market conditions to see the extent to which they are consistent with TCA or other explanations. We propose a contingency theory of entry mode performance that incorporates the effects of market and firm factors on production costs as well as transaction costs. We test our model using a nationwide sample of export market entry modes used by Australian manufacturing firms.
Our research contributes to the development of research regarding TCA and foreign market entry mode performance by addressing a number of the problems raised by Rindfleisch and Heide (1997). Specifically: (a) we incorporate the temporal nature of interfirm relations as part of transaction cost theory; (b) we show how the concept of frequency of transactions can be interpreted as a market size constraint and include this in our model and empirical analysis; (c) we examine the effects of different types of environmental uncertainty i.e. cultural and geographic distance and perceived risk, (f) we include various types of interaction effects between uncertainty, asset specificity, market size and the temporal nature of exchange relations. In addition, in our empirical analysis, we control for the problem of selection bias, that has called into question previous studies of the performance effects of different strategies in the management literature (Shaver 1998) via Heckman’s (1979) two step regression procedure. We are unaware of this procedure being used before in marketing to address such estimation problems.
First we describe the nature of entry modes in terms of the extent to which they are vertically integrated and interpret performance in terms of profitability. We then discuss the factors affecting the production and transaction costs of entry modes, which provides the basis for the contingency theory of entry mode performance. Following this, the research methodology is described and the results are presented and discussed. The implications of the research are considered in a concluding section.
Foreign Market Entry Modes
Market entry modes may be distinguished according to the extent to which the firm forward or vertically integrates into exporting activities. The least forward integrated mode is one in which a firm sells to an export agent based in its home country or to a locally based buying office of a foreign customer or distributor. In this case the firm does not perform any exporting activities itself. The next level of forward integration is when a firm takes on some of the exporting activities itself and sells to foreign based import agents, distributors or retailers. Third it can sell directly to foreign customers, especially industrial customers. Finally, it can establish its own foreign based sales and marketing operation, which represents the greatest amount of forward integration. Beyond this a firm could establish a local manufacturing operation via a licensing agreement, a joint venture or a wholly owned subsidiary. Here we focus only on exporting modes, where products are manufactured in one country and sold in another.
The term entry mode performance is used here in terms of economic performance or the profitability of a firms entry into a foreign market. Non profit market entry motives, such as resource and knowledge development or strategic moves against competitors, are assumed to be reflected in perceived profit levels or to have only a minor impact on entry mode performance in general. The profitability of an entry mode depends on total costs and revenues. Costs include both transaction and production costs. Production costs arise from the carrying out the exporting activities involved in linking producer and consumer, whether this is done within the firm or outsourced to other firms in the distribution channel. Transaction costs arise from the problem of coordinating different activities within and between firms. The central problem facing the firm is that of trading off the transaction and production costs associated with using external specialists against the transaction and production costs of performing the exporting activities within the firm (Dixon and Wilkinson 1988, Williamson 1981, 1985).
The revenues received depend on the value created for intermediate and final customers, the competitiveness of the market and the prices obtained, and the share of revenue going to different members of the export distribution channel. A variety of factors in turn affect these determinants of the revenue a firm receives. For the purposes of this research we assume that different entry modes result in comparable values being created and delivered to customers and that firms operate in competitive markets. This is line with previous research examining TCA. Under these conditions profitability depends on costs. Furthermore transaction costs include the risks and uncertainties associated with transactions between firms, including issues of potential conflict and hold up costs that affect the share of revenue received by a manufacturer, as well as opportunity costs. We do not offer specific propositions regarding factors affecting the revenue side of entry mode performance. However, in interpreting our results we consider the possibility that a revenue based explanation may account in part for some of the results.
Transaction Costs
TCA has been developed primarily by Oliver Williamson but other researchers have proposed additional considerations as they have attempted to operationalise and test TCA (Rindfleisch and Heide 1997). Transaction costs arise from the problem of coordinating the activities performed by different people, groups, departments or firms. They include the costs of finding transaction partners, negotiating agreements, and monitoring and controlling the activities of the transaction partner. They include the direct coordination costs involved in managing a relationship as well as potential opportunity costs (ibid).
TCA identifies three fundamental characteristics of transactions that affect the nature and extent of the coordination task facing the parties involved and indicates under what conditions different methods of coordination or governance modes are more efficient in handling the coordinating activities involved and hence minimise transaction costs. Three key dimensions of transactions have been identified by Williamson (1975,1985) as affecting the nature and extent of the coordination task: (a) the degree to which durable transaction specific investments are involved, i.e. asset specificity; (b) the frequency with which transactions recur; and (c) the level of uncertainty.
TCA categorises governance modes into two basic types - between firms or market governance and within firms or hierarchical governance. In addition, intermediate modes have been identified that depend on the nature of the relations formed over time between the transacting parties and which Williamson refers to as relational contracting or relational governance. To begin with we consider only markets vs governance modes. Later we incorporate relational governance issues.
TCA assumes that hierarchical governance modes are inherently more efficient and minimise transaction costs but this gain from internalising activities has to be weighed against the production cost economies that could be foregone by not using the resources and skills of other firms such as specialised marketing intermediaries.
Three arguments are made to justify the assumption that hierarchical intrafirm governance modes are more efficient coordinating mechanisms. First, firms are able to exercise greater control and to monitor performance better than is possible in markets and are thus able to detect and curb opportunism more effectively. Second, firms can provide longer-term rewards in the form of promotion and other incentives that reduce the likelihood of opportunistic behaviour. Third, the atmosphere or organisational culture tends to more closely align the interests of members of the firm. As a result of this hierarchies are preferred when the reduction in transaction costs arising outweighs the benefits of any economies of specialisation between firms.
This economic explanation has been used in variety of contexts to explain existing economic organisation structures. But, as noted at the outset of this article, the inherent normative assumption, that economic performance of different governance modes varies according to the three underlying characteristics of transactions, has received only limited research attention, especially in relation to foreign market entry modes. One example is research by Shaves (1998), who used the survival of different manufacturing entry modes over time as an indicator of performance and finds some support for TCA. But this measure of performance is suspect, as is acknowledged, because entry modes may not survive for a variety of reasons not related to economic efficiency. In the empirical study reported in this paper we use a more direct measure of performance.
We next consider each of the three dimensions of transactions identified by Williamson and examine how they affect entry mode performance. This is followed by a consideration of two additional factors affecting the economic efficiency of entry modes, the temporal and embedded nature of interfirm exchange relations and production costs and the way firm, transaction, and market environment characteristics affect these costs.
Asset Specificity
In TCA particular attention is focused on durable transaction specific investments that are termed asset specificity or idiosyncratic investments. Asset specificity could be in the form of specialised products and services, plant and equipment or expertise tailored to the needs of the exchange partner and which have limited value in other relations. Such assets mean that a firm is vulnerable to opportunistic behaviour on the part of the exchange partner and, in order to protect themselves, some safeguard mechanism must be introduced. Contracts cannot exhaustively define in advance every contingency and, furthermore, enforcement of contract provisions at a geographic and cultural distance, i.e. in foreign markets, is not cost effective (Macaulay 1963). Under these conditions TCA postulates that internal governance structure is preferable. When transaction specific assets are negligible, the most efficient form of governance is the market.
When transaction specific assets are high, markets are inefficient and vertical integration is presumed to offer superior adaptive properties. Vertical integration is able to circumvent the opportunistic problems caused by asset specificity because of the freer flow of information within a firm, the existence of a common organisational context that tend to align interests, and a formal power structure that can be used to overcome conflicts. Adaptations to unforeseen conditions can be made sequentially without the need to renegotiate agreements (Klein, Crawford and Alchian 1978; Rugman 1986; Williamson 1985). Empirical studies providing support for these contentions in terms of firms’ behaviour include Monteverde and Teece's (1982) study of automobile companies, and Anderson's (1985) study of electronics firms.
Frequency of Transactions and Market Size
There has been limited attention given to the impact of this dimension of TCA in previous empirical studies and Rindfleisch and Heide’s (1997) did not consider it in their review. One reason, we believe, for this lack of attention is that it has been interpreted narrowly in terms of the frequency of transactions, whereas it should be seen as only one dimension of a more general market size constraint. As Williamson (1985) argues: “the cost of specialized governance structures will be easier to recover for large transactions of a recurring kind” (p 65). The size and frequency of transactions are two dimensions of market size and, as Adam Smith first explained, market size limits the division of labour.
Market size limits the ability of firms to specialise, because of the costs involved in setting up and fully utilising specialised people, systems and equipment (Stigler 1951). This principle applies to transaction costs as well as production and costs (Dixon and Wilkinson 1986). The size of the market represented by an exchange partner depends on the size and frequency of transactions with them and this limits the division of labour with respect to the coordinating activities in the trading relationship. More specialised governance modes involve using more specialised labour and capital inputs such as partner specific electronic control and reporting systems and/or key account managers. If the amount of business conducted with an exchange partner is limited in terms of either transaction size or frequency the use of such relationship specific inputs could be inefficient because they would not be fully utilized. As Florence (1933) pointed out many years ago, economies of specialization are only potential economies because they depend on firms being able to fully utilize specialist inputs which have various efficient scales of operation.
The size and frequency of transactions between exchange partners leads to what are known as economies of bulk transactions (Florence 1933, Dixon and Wilkinson 1985). Transaction costs do not rise in direction proportion to the size of a transaction because the time to negotiate, monitor and control an exchange does not increase in direct proportion to the amount of business involved. Communication and transport cost economies arise and specialised people and resources can be more fully utilised. Therefore, in terms of choice of entry mode, the greater the amount and/or frequency of transactions or, more generally, the size of a foreign market, the more a firm is able to fully utilize a specialised governance structure such as a management hierarchy or some form of relational contracting.
Williamson (1985) also points to an interaction effect between asset specificity and the market size of an exchange relation. When asset specificity is low market contracting can be used for both low and high frequency transactions. But when asset specificity is high vertical integration becomes more likely as the amount and/or frequency of transactions increases. This is because the size of the market becomes big enough to support a specialised internal governance structure.
In addition to the market size of an exchange relation the overall size of the firm affects its transaction costs (Nooteboom 1999). Search costs are higher for smaller firms due to lack of specialised staff and resources and this “makes the set-up costs of governance expensive relative to the size of the transaction” (p20). This reduces the efficiency of vertically integrated entry modes for smaller firms.
Uncertainty
Uncertainty arises as a result of a firm’s bounded rationality, which limits its ability to anticipate all the consequences of its actions and hence to write comprehensive contracts. Two types of uncertainty are distinguished in TCA. Environmental uncertainty relates to the risks associated with entering markets and cultures with which a firm is less familiar and comfortable operating in. This makes any exchange agreement difficult to specify in advance because conditions could change from the time the agreement was reached. Behavioural uncertainty is transaction partner focused and concerns the problem of determining whether an exchange agreement has been adhered to, either because performance is intrinsically difficult to measure or because of the high costs of getting the appropriate information, e.g. when a firm has to assess the technical service performance of an intermediary. Behavioural uncertainty depends on the nature of the product and service involved. The more complicated or technology and knowledge intensive a product is, the greater is the likely behavioural uncertainty and the more costly it is to use ordinary markets (Anderson and Gatignon 1986, Teece 1986).