A Hierarchical Model of Export Entry Mode Choice
Ian F. Wilkinson, School of Marketing, University of New South Wales Australia
Paper presented at Australian New Zealand Marketing Academy Annual Conference
Deakin University, Melbourne December 2002
Introduction
Export entry modes can be classified in terms of a two stage hierarchy as shown (a) whether they are vertically
integrated (VI) or not (NVI) and (b) the particular form of VI or NVI mode used. VI modes are those in which a
firm performs all or most exporting activities itself and include direct exporting to foreign customers and
establishing a foreign sales branch as alternative forms. NVI modes are those in which a firm uses specialist
intermediaries to carry out exporting activities on its behalf and include domestic based export agents and
foreign market based import agents and distributors as alternative forms.
Previous studies of entry mode choice tend to obscure the differential impact of factors at each stage of this
hierarchy because of the way the analysis is undertaken. Typically a set of entry modes is compared involving a
mix of VI and NVI forms and binomial or multinomial logit is used to identify factors associated with the use of
each mode. However, the factors distinguishing between VI and NVI modes are not likely to be the same as
those distinguishing between different forms of VI and NVI modes. Hence the interpretation of such results is
ambiguous and factors explaining the choice between particular forms of VI and NVI modes may appear to be
insignificant.
The purpose of this research is to examine the extent to which different factors impact on each stage of export
entry mode choice. In part it builds on research by Pan and Tse (2000) who tested a hierarchical model of direct
foreign investment and showed that different factors affect choice at each stage.
Theories of Entry Mode Choice
TCA presumes that formal hierarchies are more efficient than markets because of the ability to exercise greater
control, to monitor performance better, to provide longer term rewards and because the atmosphere within firms
more closely aligns people’s interests. A firm has to trade off the potential gains of lower transaction costs
resulting from internalising exporting activities against the potential gains of using a specialist intermediary that
may have lower production costs in performing the relevant activities. The general hypothesis arising from this
is that the greater the transaction costs the more likely are firms to use vertically integrated export entry modes.
The network approach (e.g. Axelsson and Easton 1992, Ford 1997, Hakanasson and Snehota 1995, Johanson
and Mattson 1994) focuses on the way transactions are embedded in a social and normative context
(Granovetter 1985). Transactions are parts of on-going relations between economic actors that have a past and a
future and can involve both economic and social dimensions. The development of such relations alters the
nature of the coordination problem facing firms, including the type of uncertainty and likelyhood of
opportunistic action by counterparts. As a result market relations become a potentially efficient and desireable
means of coordination that sit between the extremes of formal hierarchies and arms length markets. This leads
to the second general hypothesis, that the more developed the relations between economic actors the less likely
are firms to use vertically integrated export entry modes.
Williamson (1975,1985) identifies three key dimensions of transactions that affect transaction costs and hence
the choice of governance mode: (a) asset specificity; (b) transaction frequency and size; and (c) uncertainty. We
consider each in turn and how they relate to TCA and NWA.
Asset Specificity
Durable assets such as specialised products and services, knowledge, plant and equipment that have limited
value in other relations accentuate the problem of opportunism for a firm. TCA argues that this increases the
preference for internal governance. In NWA asset specificity is something that is co-produced in a relationship
over time and is thus embedded in it (Badaracco 1991, Hakansson 1987a 1987b). Such are less susceptible to
opportunistic action as they are co-produced and “owned.”
2
Frequency of Transactions and Market Size
This dimension has been given limited attention in TCA (Rindfleisch and Heide 1997) because it has been
interpreted too narrowly. Market size limits the division of labour and this, applies to both production and
transaction costs (Dixon and Wilkinson 1986. The size of the market with an exchange partner depends on the
size and frequency of transactions (Williamson 1985), as well as the duration of the exchange relation, and this
limits the ability to utilise more specialised governance modes. The size of the firm also makes the costs of
establishing more specialised governance expensive relative to transaction size (Nooteboom 1999). Thus we
expect market and firm size and the duration of the relation to limit the use of VI modes.
Uncertainty
Uncertainty in international markets stems from firms operating in unfamiliar market environments and from the
problems of determining whether an exchange agreement has been adhered to. These in turn depend on how
complicated or technology intensive a product (Anderson and Gatignon 1986, Teece 1986) and on the
geographic and cultural distances involved (Johanson and Vahlne 1977). Psychic or cultural distance varies
according to the international business knowledge and experience of firms, which in turn relates to the extent of
their internationalisation and experience in a particular market (Johanson and Vahlne 1977, 1990, Petersen and
Pedersen 1997). TCA argues that the greater the uncertainty the more likely are firms to use VI modes. NWA
leads to a different hypothesis. Uncertainty due to lack of familiarity with a market leads to the use of
knowledgeable specialist intermediaries until, over time, psychic distance is reduced as the firm becomes more
involved and familiar with a market. Others have also argued that increased leads to the use of independent
intermediaries because it is quicker and less costly to change organisation structures compared to a VI structure.
(Klein et al 1990, Shelanski and Klein 1995, Rindfleisch and Heide’s 1997).
Production Costs and the Economies of Scale and Scope
The potential economies of scale and scope available to a firm to perform (i.e. produce) activities internally,
stem from the scale and scope of its own operations. The larger its scale in performing an activity the more a
firm can gain scale and scope economies internally and hence the less likely it is to use external specialists
(Dixon and Wilkinson 1986). Various dimensions of a firm’s operations affect the scale and scope economies
available to it including firm size, level of internationalisation and foreign market involvement.
Revenue Effects and market size
Additional types of explanatory variables have been proposed including country and industry specific
institutional factors, the effects of tariffs and other trade barriers. In the empirical research a dummy
variable is used to control for type of industry but other types of factors are not included.
H8: The industry context in which a firm operates will affect how likely firms are to use more
vertically integrated market entry modes
Research Methodology
Sampling
The data used to test the hyptheses comes from a comprehensive survey of Austrlian manufacturers carried out
in the early 1990s. Questionnaires were mailed to all firms on to a list of Australian manufacturing firms
supplied by Dun and Bradstreet, including both exporters and non exporters, except in the case of small nonexporters (under 100 employees) where a systematic sample of one in four firms were mailed a questionnaire.
A reminder letter was sent after two weeks to try to improve the response rate.
From the 3942 questionnaires mailed a total of 550 useable responses (i.e. a 14% response rate) were received
comprising 403 exporters and 147 non exporters. Only the exporters are of interest here and the response rate
for them was 19%, twice that for the non-exporting firms. When only exporting firms are considered, the
3
achieved response rate falls within the range of other published survey research on firms, that range from 14%
(Phillips 1981) to 49% (Cox et al. 1974). A comparison of early and late responding firms indicated no
significant differences except for firm size. There was a greater percentage of larger firms among those
responding later, possibly indicating the time it took to gain approval for participation and for completing the
questionnaire.
One reason for the modest response rate may be the length of the questionnaire. The issue of questionnaire
length has been the subject of considerable investigation but with often confusing and conflicting results
(Childers and Fereel 1979). Another factor affecting response rates was the state of the economy at the time of
the study. It was conducted during an economic depression in Australia including a high inflation rate and a high
level of unemployment. Several firms wrote back to say that in these economic conditions they did not have the
time to complete such a long questionnaire. Given the objectives of this study are to examine the impact of
different factors on choice of entry mode the sample was considered appropriate for this purpose. For the
purposes of analysis one export country was chosen from among those reported by a firm. The procedure used
is explained in the next section.
Measures
a) Export Entry Mode.
Firms were asked to nominate up to five countries they exported to in order of importance and to describe their
entry mode in terms of the following types.
1.
2.
3.
4.
5.
6.
7.
Overseas branch, sales subsidiary or associate company
Australian export agent, distributor or merchant
Foreign firm’s Australian buying office
Direct to overseas wholesaler, distributor or agent
Direct to overseas retailer
Direct to overseas end user
Other (write in)
A total of 1231 cases of country market entry were described, which were summarised in terms of four types of
entry modes as follows:
1. The firm sells its products to foreign buying offices or Australian export distributors in Australia i.e.
indirect exporting.
2. The firm exports directly to distributors,or retailers in the foreign market
3 The firm exports directly to end users in the foreign market,
4 The firm exports to its branches or sales offices in a foreign country, which then sells their
products to customers in those markets.
Other types of entry modes, such as franchising, licensing and manufacturing abroad through joint ventures or
wholly owned subsidiaries are not considered. The classification of modes is designed to reflect different levels
of vertical integration, vary from 1, sales via Australian based intermediaries, which is the least amount of
vertical integration to 4, exporting using the firms foreign branches or sales offices and is similar to those used
in other studies of entry modes. This resulted in a sample of 1088 cases using one of the four types of entry
modes made up of 9% via an Australian based intermediary, 50% via a foreign market base intermediary, 25%
direct to users and 14% via the firm’s foreign branch office. 100 firms reported using more than one of the four
entry modes in the same market, which raises interesting issues for theory development that have not been much
considered in the literature (Pedersen and Welch 2002). Here, these cases were excluded from analysis as did
Klein et al (1990).
The sample of market entry modes is not completely independent because firms could be exporting to more than
one country and this could bias the results. One export market was chose at random from those described by a
firm. In previous the main export country has been used as the basis for analysis but a differerent method was
used here for two reasons. First, selecting the main export market tends to reduce the variance in the data as it
biases the sample towards more important export markets. Secondly the distribution of entry modes is skewed
towards the use of foreign market intermediaries because this is the most commonly use method of entry used
by Australian firms. In order to improve the distribution of entry modes in the analysis sample and to maximise
4
variance in export market conditions we randomly selected the export market from among those described by a
firm which used one of the four types of entry modes. In order to reduce the skew in the distribution of entry
modes in the sample, the random number was weighted by the type of entry mode used i.e. 1 minus the
probability of the entry mode reported in the full sample set i.e (1-0.09) = 0.9 for Australian based
intermediaries, (1-0.50) = 0.5 for foreign market intermediaries, (1-0.25) = 0.75 for direct to users and (1-0.16)=
0.84 for foreign branches. This resulted in an analysis sample of 343 market entries, of which 11.7% were via
an Australian based intermediary, 39.9% a foreign market intermediary, 28.3% direct to user and 20.2% via a
foreign branch office.
b) Product Asset Specificity. Asset specificity is treated as a latent construct which was measured using multiitem scales derived from respondent’s ratings of the product that accounts for the greatest percentage of their
firm’s export in terms of the eight attributes shown in Table 2. A four point scale ranging ranging from 1=
strongly disagree to 4 = strongly agree was used. The statements reflect the extent to which the products
requires special training and support, is unique and differentiated and employs specialised technology. They
were based on items used as part of the Company Readiness to Export product attribute questions (Cavusgil and
Nason 1990). Exploratory factor analysis using maximum likelihood estimations was used to identify
underlying dimensions and this indicated that were three significant factors, which account for 66% of variance.
The first dimension reflects advanced technology and uniqueness of the product and is labelled “Hitech,” the
second reflects the extent of pre and post sales support required and is labelled “Service.” The third, refects the
degree of standardisation and similarity to domestis product. Because the alpha for the items comprising this
scale is only .46 this dimension is dropped from further analysis. Hitech and Service represent different
dimensions of Product Asset Specificity and the scales constructed using the high loading items have acceptable
alphas as shown in Table 2.
c) Foreign Market Scale and Scope: Two measures were used
Foreign Market Size: Measures were obtained of the percentage of a firm’s total exports to a
market, the percentage of its total turnover accounted for by exports and the firm’s total turnover
(total sales) in the previous year (rated on a seven point scale from 1= up to $100,000 to 7 = more
than $20,000,000). Previous research indicated that firms were reluctant to report actual dollar
turnover figures. From this we estimated the volume of business conducted in a country by
multiplying the three percentages together and multiplying them by the midpoint value of their
rating of their firm’s total turnover (for ratings of more the $20 million an estimate of $50million
was used). This procedure is similar to that used by Klein et al.(1990).
Years a firm has been exporting to the foreign market.
d) Uncertainty. Uncertainty was measured in terms of cultural and geographic distance and management’s
perceived risk of exporting
Cultural distance. Measured using a separate questionnaire that was distributed to 100 members of
staff and postgraduate students at two universities in Sydney. They were asked to rate various
countries in terms of their socio-cultural distance from Australia on a 5 point scale from 1 = very close
to 5= very distant. A response rate of 47 percent was obtained. On the basis of these results countries
were classified into three groups, low, medium and high as shown in Table 2.
Geographic distance. Measured using Airline distances between capital cities and Sydney. These
were classified into 3 groups: close (less than 5000 kms), medium 5000 to 10,000kms) and distant
(greater than 10,000kms) as shown in Table 2.
Perceived Risk. In addition to country distance uncertainty is measured in terms of a multi-item
measure of a respondent’s degree of agreement with 5 statements about the risks of international
operations using a 4 point scale from 1= strongly disagree to 4 = strongly agree. The four items used
are shown in Table 3. The statements were based on previous studies of exporting (Dunphy and Layton
1969, Barrett and Wilkinson 1985, 1986). The alpha for the scale is 0.88.
e) Firm Scale and Scope This was measured in terms of the firm’s overall size and in terms of various
measures of the scale and scope of its exporting operations.
5
Firm Size. This is measured on a five point scale from 1= under 10, 2= 10-19, 3=20-49, 4=50-99, 5=
100-500 and 6=500+. In order to convert this ordinal scale to one with more interval scale properties,
the scale points were recoded to the midpoints of their respective size range, with a value of 750 used
for the 500 and over category.
Scale and Scope of a Firm’s Exporting Operations. four indicators were used.
 percentage of a firm’s turnover accounted for by exports
 years the firm had been involved in exporting.
 number of countries exported to
 number of different products exported.
f) Type of product. Firms were asked to describe the main product they exported and this was classified into
the standard international classification of products. This showed a spread of types of products which were
grouped into consumer, industrial or mixed products. Dummy variables were constructed for consumer
products and industrial products.
Results
Predicting Entry Mode Choice
Multinomial logistic regression is used to test the extent to which the various explanatory variables account for
the use of each of the four types of entry mode. This approach is similar to that used in other studies of entry
mode (e.g Klein et al 1990, Meyer 2001). In the analysis a foreign branch is the mode of entry used as the
default in the regression. This means that the regression coefficients indicate the impact of the variable on the
probability of a firm using a particular mode compared to using a foreign branch. Stronger coefficients for one
mode compared to another indicate that the variable has a larger impact on the probability of using that mode
compared to others. We use a 10% level of significance throughout the analysis to guide interpretation, as we
believe this to be more appropriate at this stage of development and testing of theory in the area. In addition,
various environmental factors such as tariffs and other forms of regulatory controls may distort and suppress the
impact of the factors considered in our model as the measures used are only indirect measures of production and
transaction costs. However, we will treat with more caution any results that are only significant at the 10%
level.
Two kinds of models were estimated. First, a model including only the main effects is estimated, using the
measures of asset specificity, uncertainty and foreign market involvement and firm scale and scope, plus dummy
variables for consumer and industrial products. The second model includes various types of interaction effects
described in the hypotheses.
a) Main Effects Model
Table 4 shows the results of the main effects model. The model is significant in terms of the chi-square test and
Pearson goodness of fit. It correctly predicts 52.3% of the modes used and the Cox and Snell pseudo R-square
is 0.308 The most significant predictors of mode used relative to foreign branches are cultural distance and firm
size, indicating that these modes are more likely to be used than foreign branch offices when cultural distance is
larger and when firms are smaller. In addition to these, the service measure of asset specificity (inversely) and
the number of countries exported to (inversely) are both significant predictors of the use of domestic export
intermediaries. Lastly, geographic distance (inversely) and number of products exported (inversely) are
significant predictors of the use of direct exporting to users relative to foreign branches.
b) Interaction Effects Models
As (Rindfleisch and Heide 1997) note, main effects can be confused with interaction effects when interaction
effects are not included in the model tested. However, a potential problem of multicollinearity arises because
the interaction terms can be highly correlated to the main effect measures of which they are composed. This
tends to increase standard errors and make both the main and interaction effect non-significant (Rindfleisch and
6
Heide 1997). In addition interaction terms including a common component measure, such as the interaction
between uncertainty and asset specificity and the interaction between uncertainty and years in a market, can be
highly correlated if the variance of one measure dominates the interaction term. The correlation matrix, together
with the means and standard deviations of the measures is shown in the appendix.
In order to control for these types of problems we examined two models with interaction effects. The first
model excluded all main effects that formed part of any interaction term. In addition we inspected the
correlation matrix and excluded one of a pair of interaction terms if the correlated at 0.75 or greater. At a
correlation of 0.75 more than half the variance is shared by the two terms. In choosing which measure to
exclude we tried to ensure at least one measure of each type of interaction was included in the model. The
resulting model (not shown) proved weaker than the main effects model in terms of percent of correct
predictions (51.9% versus 52.3%) and the chi-square change from including an additional 9 variables in the
model was not significant compared to the main effects mode. i.e. 113.55 – 104.14 = 9.41. Only the size
variable was a significant predictor of all modes and the number of products exported and the interaction of
cultural distance with years in exporting were significant predictors of direct exporting. Hence concluded that
the interaction terms did not fit the data better than the main effects.
Hence a second model was estimated in which the significant main effects were included in the model together
with the interaction terms. Once again we excluded any interaction term that correlated at 0.75 or greater with
a main effects measure or deleted one pair of any two interaction terms that correlated at 0.75 or greater. This
resulted in the exclusion of 7 interaction terms i.e. cultural distance with the hitech measure of asset specificity,
perceived risk with the service measure of asset specificity and with years exporting to the foreign market,
perceived risk with years exporting to the foreign market and with years exporting, sales to the foreign market
with the hitech measure of asset specificity, and sales to the foreign market with years exporting.
The results for the second model are shown in Table 5. Only one of the interaction terms is a significant
predictor. The interaction between cultural distance and the years a firm has been exporting increases the
probability of direct exporting compared to establishing a foreign sales branch. The overall model is significant
but the improvement over the main effects model is not significant. Hence there is limited support for the
impact of interaction effects.
One reason for the lack of significance of the interaction effects is the entry mode decision is a hierarchical
decision in which different factors play a role at each stage because of the relative importance of transaction
costs versus production costs for example. This could suppress the significance of some factors when all four
modes are considered. Such a model of entry mode choice has been proposed by Pan and Tse (2000). Their
model distinguishes two stages. At stage one the choice is between equity and non-equity modes and at a
second stage different types of equity and non-equity modes are selected. In terms of the entry modes
considered here we may distinguish between those modes that involve use of a specialist external intermediary,
i.e. a domestic or foreign based agent or distributor, and modes that involve the firm internalising the exporting
activities by exporting directly or by establishing its foreign branch office. Having decided whether they will
internalise the exporting activity or not the firm then has to decide what form this will take.
In order to test this two stage model we estimated three dichotomous logit models. The first stage model
focused on the probability of internalising versus externalising the exporting function. The two second stage
models focused on (a) the probability of choosing a foreign versus domestic based intermediary and (b) the
probability of choosing a foreign branch office over direct exporting. Stepwise logistic regression was used to
estimate the models, with entry parameters set at 15% to ensure that all potential predictor variables were
considered, but only models where the predictor variables were significant at the 10% level or better were
considered. We also checked for multicollinearity among the variables entering the model. The results are
shown in Table 6.
The first stage model is significant and shows that geographic distance is inversely related to use of internalised
modes and size as directly related to use of internalised modes. Two interaction terms are significant. That
between geographic distance and the service dimension of asset specificity is directly related to internalising
export activities and that between cultural distance and years a firm has been involved in exporting is also
directly related to internalising export activities.
Turning to the second stage models, the choice between domestic and foreign based intermediaries is not well
explained by the predictor variables. The interaction of cultural distance and years in exporting is marginally
7
significant and indicated that foreign based intermediaries are less likely to be used when this is greater. The
probability of using a foreign branch compared to direct exporting increases with cultural distance, the size of
the firm and the number of products exported. The interaction of geographic distance and the service dimension
of asset specificity increases the likelihood of establishing a foreign branch office.
Discussion
To simplify the discussion we will focus only on the main effects multinomial model and the hierarchical model
results. First we consider the direct effects of each type of predictor variable and then we consider the
interaction effects.
Only the service measure of asset specificity has a significant impact on entry mode choice in the case of the
multinomial model. Establishing a local branch office compared to using domestic export agents is more likely
when products require higher levels of service. This supports hypothesis H1 in that high levels of before and
after sales service lead to the development of human asset specificity. Sales and service personnel interact more
closely with customers and have to develop a closer understanding of them. The result is consistent with other
research that has focused on human asset specificity (Rindfleisch and Heide 1997).
The absence of any significant relationship with the Hitech measure suggests that this is not a transaction
specific asset but more a firm specific asset that helps it differentiate its offer. Because only exporting modes
are being considered, there is a smaller risk of opportunistic behaviour from foreign counterparts appropriating
the firm’s production technology. This dimension is likely to be more important when foreign manufacturing
modes are considered such as licensing or joint ventures compared to a fully owned operation.
Sales to the foreign market or duration of involvement in the market are not significant predictors of entry mode
in either the multinomial or hierarchical models, hence there is no support for hypotheses H2 and H3.
Cultural distance has a strong link with choice of entry mode in the multinomial model. The less integrated
entry modes are all more likely to be used compared to foreign owned branches when cultural distance and
hence uncertainty is greater. This does not support the TCA hypothesis H7 but does support the NWA
hypotheses H9. Firms are reluctant to set up foreign branch offices when cultural distance is larger because this
exposes them to more risk than other modes. The effect is stronger for direct exports compared to using
domestic or foreign market based intermediaries, indicating that direct exporting is more sensitive to cultural
distance than the other two modes.
Further clarification comes from the second stage of the hierarchical model. This shows that cultural distance is
not a significant predictor of the choice between internalising and externalising exporting activities but it is a
significant predictor of the choice between the two forms of vertically integrated modes i.e. direct exporting
versus a foreign branch office. Given the choice of a vertically integrated mode, the greater the cultural distance
the more likely are firms to use direct exporting rather than a foreign branch office. Again the result does not
support TCA hypothesis H7 and does support H9.
The multinomial model indicates that firms are less likely to use direct exports compared to foreign owned
branches when geographic distance is greater. Firms serving markets very distant from Australia, such as
Europe and the USA, are more likely to set up local branch offices than to serve them directly from Australia
because of the problems of communication, which can be a source of behavioral uncertainty. This result lends
some support to hypothesis H7 and not H9 in terms of the choice between types of vertically integrated modes.
The results for the hierarchical model indicate that geographic distance is associated with using specialist
intermediaries and not vertically integration, which is contrary to H7 and supports H9.
Both models indicate that larger firms are likely to use more vertically integrated modes. They are more likely to
use foreign branch operations compared to the other three modes, are less likely to use specialist intermediaries
and are more likely to prefer a foreign sales branch to direct exporting. All these results support hypothesis H5.
The effect of scale and scope economies is also reflected in the significance of measures of a firm’s
internationalisation. Firms who export to fewer countries are more likely to use domestic based agents rather
than to establish foreign branch offices and firms who export fewer products are more likely to export direct
than to establish a foreign branch office. These results support hypothesis H11.
8
Turning now to the effect of interaction terms, we focus on the hierarchical model. The service dimension of
the product appears to sensitise the firm to the effect of geographic distance and increases the use of vertically
integrated modes compared to using specialist intermediaries. The problems of delivering adequate service
levels in more distant markets encourages firms to forward integrate into exporting. This supports hypothesis
H8. The correlation between geographic distance and the interaction term is 0.60 and hence the difference in
direction of results for the main effect of geographic distance and the interaction effect with years in exporting is
not likely to be a result of multicollinearity.
Given that firms are using a vertically integrated mode, the interaction of geographic distance and the service
dimension of the product also affects the second stage choice between direct exporting and foreign branch
offices. In other words, the effect of geographic distance depends in part on the type of product being exported.
Firms are more likely to establish a local branch office when geographic distance and the service component of
a product is high, which supports hypothesis H8.
The interaction of cultural distance and years in exporting makes firms more inclined to internalise export
activities rather than relying on specialist intermediaries. The effect of export experience moderates the effect
of cultural distance and makes firms more willing to forward integrate into exporting. Interestingly this
contrasts with the multinomial main effects result indicating that firms are more likely to use any of the less
integrated modes compared to foreign branch offices, when cultural distance is high. The effect of export
experience thus moderates the effect of cultural distance on uncertainty. Experienced firms have more
knowledge of foreign and dissimilar markets and are able to perform more of the export activities themselves
because internal production costs are now lower or as a means of reducing transaction costs. Hence the results
provide support for H10.
The interaction of cultural distance and years in exporting is also a marginally significant predictor of the choice
between domestic and foreign based intermediaries. The result suggests that export experience reduces the
probability of using foreign based intermediaries when uncertainty is high. This seems contrary to the preceding
result but a possibly explanation is that when uncertainty is high more experienced firms prefer to deal with an
export intermediary based in the domestic market, with whom they can have more direct and frequent contact,
rather than a foreign based one. The effect of email, faster travel and the internet may tend to diminish this
effect over time. Not too much should be made of this results as it is only marginally significant (p = 0.64).
The lack of significance of the interaction terms between uncertainty and the duration of a firm’s involvement in
a particular market is interesting. It appears that the experience gained generally in international markets is
more important in affecting how firms deal with uncertainty and select entry modes, rather than the experience
gained in a focal market. The role and importance of developing relations with intermediaries can be learned in
any market and then influences behaviour in other markets with the same kinds of uncertainty. In addition the
same intermediary firm may be used to handle exports in a number of markets. The resources and expertise
gained from international operations generally, rather than just in the focal market, impact on the way a firm
moves up the establishment chain in the focal market.
Conclusions and Future Research Directions
Our review of the transaction costs and network approaches to understanding entry mode choice and
performance identified a number of competing as well as complementary hypotheses regarding the nature and
causes of asset specificity and its impact, the nature, sources and role of uncertainty, and the effect of the
temporal nature of interfirm relations. In addition, several potential interaction effects were identified among
these dimensions.
Our results show that support exists for hypotheses stemming from both the transaction cost and network
approaches in terms of factors affecting coordination costs between firms. They also show that factors affecting
production costs, ie the costs of carrying out exporting activities within a firm as opposed to being carried out by
a specialist intermediary, also have a strong impact on entry mode choice and performance. An overview of the
results is presented in Table 10, which shows the hypotheses associated with each type of predictor variable and
the results supporting or not particular hypotheses.
Our results underscore the need to consider both transaction costs and production costs understanding entry
mode choice and performance. They also reinforce the importance of the temporal nature of firm relations and
that the efficiency of individual transactions cannot be considered in isolation because of the development of
9
relations and firm learning. Over time. Significant interaction effects have been found between several types of
factors that need to be confirmed in future research. These include interactions between the asset specificity of a
product, in particular its service dimension, and uncertainty arising from cultural and geographic distance and
between uncertainty and a firm’s experience in a market and in exporting generally. exporting and the role
played by uncertainty and the size of its market involvement.
The choice of export entry mode is shown to be more complex than a choice between alternative governance
modes to reduce transaction costs. A two stage hierarchical model of entry mode has been proposed in which
different factors are important at each stage, and is based on a similar model developed by Pan and Tse (2000)
of equity versus non-equity modes of foreign market entry..
At the first stage is the choice between using an external specialist intermediary of some kind or integrating
forward into exporting activities. This depends critically on production costs i.e. the ability to gain economies
of scale and scope internally in carrying out the various types of international marketing activities involved..
This in turn depends on the firm’s overall size and the geographic distance to the market. Larger firms are more
able to forward integrate into exporting but geographic distance makes coordination more difficult hence limits
forward integration. An additional factor is the type of product involved which affects the impact of geographic
distance. For products involving high levels of service, firms are more likely to forward integrate when
geographic distance is greater, because of the problems of providing adequate service at a distance. Lastly, the
impact of cultural distance on use of specialist intermediaries is moderated by the experience of the firm in
exporting. More experienced exporters confronting greater cultural distance are more likely to forward integrate
into exporting than less experienced exporters, presumably because the perceived or “psychic” distance is lower.
If production cost considerations or geographic distance prevent a firm from internalising the exporting
activities, the choice is then between a local or foreign based intermediary. This choice is not well explained by
our model but seems to depend in part on uncertainty and the experience of the firm in export markets. Firms
with more international experience are more likely to use a domestic based intermediary when cultural distance
is larger, which may reflect their desire to have closer and more frequent contact.
If a firm can efficiently undertake its own exporting activities it can choose between exporting direct to
customers or establishing its own foreign sales operation. This depends on size, uncertainty and type of product.
Larger firms that export a number of different types of products are more likely to set up a local branch office.
But the problems due to cultural distance lead firms to prefer exporting direct to users rather than incurring the
extra costs and risks of setting up a local office. Lastly, when the geographic distance to the market is large and
the product requires higher levels of service, firms are more likely to set up a local branch office.
As with any research of this kind there are a number of limitations to be acknowledged. First, it is based on a
study of manufacturing firms only carried out in a one country. Some of the results may reflect the peculiarities
of Australia’s location, its export markets and the composition of its exports. In particular, it is a primary
industry based economy with a focus on exports of simply rather than elaborately transformed manufactures.
Second, we focused on exporting and did not include foreign manufacturing as an entry mode. Third, the study
did not attempt to measure transaction or production costs directly but only the factors theory suggests impact
on these costs. Fourth, we did not include various legal and institutional factors constraining choice of entry
mode in particular countries.
Based on the experience of this research some improvements can be suggested for future research of this kind.
Better measures of some of the constructs are required. First, direct measures of transaction and production
costs could be used, which would involve gathering more information about the costs and history of particular
export entry modes and the nature of the relations a firm has with foreign counterparts. A more direct measure
of product asset specificity is required as well as relationship asset specificity. We focused on the service nature
of products and the resultant human asset specificity required, as other researchers have done (Rindfleisch and
Heide 1997), but asset specificity includes the commitment of other resources. It is also important to distinguish
between asset specificity that exists prior to a relation and asset specificity that is co-produced as a result of a
relationship. In this study we used the length of time a firm has been involved in a market and in exporting
generally as indicators of a firm’s likelihood of developing closer exchange relations and therefore relationship
specific assets. More direct measures are required of a firm’s relationship resources including relations with
intermediaries or customers it deals with in more than one market. The symmetric or asymmetric nature of any
relationship specific assets between the exchange partners should also be measured. Lastly, the measures of
10
cultural and geographic distance could be refined to reflect the perceptions of individual decision makers rather
than a general characteristic of a market.
Finally, we offer some additional suggestions for future research directions. First, alternative modelling methods
should be used to examine the causal paths among key constructs rather than using logit to regress entry modes
on the various predictor variables. This would allow performance measures as well as more direct measures of
production and transaction costs to be included in the model. In addition to cross sectional survey based
methodologies we also need more studies of the development of entry modes over time and more detailed
activity based costing analysis of particular entry modes under different conditions.
Second, our research revealed a number of cases in which firm’s use multiple entry modes in the same market.
Other researchers have reported similar findings. Petersen and Welch (2002) have focused attention on this
area and suggested some explanations but there is a need for theory development in the area of combination
entry modes as well as case studies and model testing.
Third, the network approach suggests the need to include additional types of explanatory variables into our
models of entry mode choice (Wilkinson 2001, Wilkinson and Young 2002). This includes the characteristics
of the relations a firm has with other firms such as suppliers (especially international suppliers), intermediaries,
customers, competitors and complementary firms. These relations may affect choice of country and entry mode,
as when a firm follows actual or potential domestic customers into international markets (???Japanese study, or
when it learns from its input relations about export opportunities Welch and Welch input-output rels???. As we
have already noted, the characteristics of the domestic and international market networks in which a firm
operates are likely to have significant impacts on entry mode choice and performance. These factors need to be
incorporated into future models and empirical research..
Bibliography
Acheson J M, 1985, The Main Lobster Market: Between Market and Hierarchy, Journal of
Law, Economics, and Organization, 1(2), pp. 385-398
Achrol, Ravi S. (1997), "Changes in the Theory of Interorganizational Relations in Marketing: Toward a
Network Paradigm.” Journal of the Academy of Marketing Science, 25 (1), 56-71.
Achrol R S, T Reve, and L B Stern, 1983, The Environment of Marketing Channel Dyads: A
Framework For Comparative Analysis, Journal of Marketing, Vol 47, Fall, pp 55-67
Ahmed, Farid, Paul Patterson and Chris Styles (1999) “The Determinants of Successful Relationships in
International Business” Australasian Marketing Journal, 7(1) 5-21
Aldrich, John H. and Forrest D. Nelson (1984) Linear Probability, Logit, and Probit Models,
Beverly Hills: CA, Sage Publications
Anderson E, 1985, The Salesperson as Outside Agent or Employee: A Transaction Cost
Analysis, Marketing Science, 4, Summer, pp. 234-254
Anderson E. and H. Gatignon, 1986, Modes of Foreign Entry: A Transaction Cost Analysis
and Propositions, Journal of International Business Studies, XVII, 3, Fall, pp. 1-26.
Anderson, C., Hakan Hakansson, and Jan Johanson (1994), “Dyadic Business Relationships
Within a Business Network Context.” Journal of Marketing, 58, (October), 1-15
Axelrod R M, 1984, The Evolution of Cooperation, New York: Basic Books
Axelsson, Bjorn and Geoff Easton eds.(1992) Industrial Networks: A New View of Reality,
London, Routledge
11
Badaracco J. L. Jr. (1991) The Knowledge Link: How Firms Compete Through Strategic Alliances. Boston:
Harvard Business School Press
Barrett N J and I F Wilkinson, 1986, Internationalization Behavior; Management Characteristicss of Australian
Manufacturing Firms by Level of International Development, in Turbull P W and S J Paliwoda (eds),
Research in International Marketing, Croom Helm, USA
Bilkey W.J, 1987 Profitable Export Marketing Practices: An Exploratory Inquiry, in Rosson P.J. and S.D. Reid,
(eds.), Managing Export Market Entry and Expansion: Concepts and Practice, New York: Praeger
Bilkey W.J. and G. Tesar, 1977, The Export Behaviour of Smaller-sized Wisconsin Manufacturing Firms,
Journal of International Business Studies, 8, Spring/Summer, pp. 93-98
Blankenburg-Holm, Desiree, Kent Eriksson and Jan Johanson (1996), “Business Networks and Cooperation in
International Business Relationships," Journal of International Business Studies, 27 (5), 1033-1053
Buckley, Peter J. and Marc C. Casson (1998) “Analyzing Foreign Market Entry Strategies: Extending the
Internalization Approach” Journal of International Business Studies 29 (3) 539-561
Cavusgil S. Tamer andRobert W. Nason (1990) “Assessment of Company Readiness to Export” in Hans
Thorelli and S. Tamer Cavusgil International Marketing Strategy, 3 rd edition, 129-139
Childers T L and O C Fereel, 1979, Response Rates and Perceived Questionnaire Length in Mail Survey,
Journal of Marketing Research, XVI, pp.429-431
Contractor F.J., 1984, Choosing Between Direct Investment And Licensing: Theoretical Considerations And
Empirical Tests, Journal of International Business Studies, Winter pp. 167-188
Cox E.P., W.P. Anderson and D.G. Fulcher, 1974, Reappraising Mail Survey Response Rates, Journal of
Marketing Research, Vol.11, Nov., pp. 413-7
Donaldson, Lex (2001) The Contingency Theory of Organizations, Thousand Oaks, CA: Sage
Dixon D F and I F Wilkinson, 1986, Toward A Theory of Channel Structure, in Sheith (ed), Research in
Marketing, vol 8, JAI Press pp. 27-70
Dwyer F R, P H Schurr and S Oh, 1987, Developing Buyer-Seller Relations, Journal of Marketing, 51, April,
pp. 11-28.
Dyer, Jeffrey H. and Harbir Singh (1998) “The relational View: Cooperative Strategy and Sources of
Interorganizational Competitive Advantage” Academy of Management Review, 23 (4) 660-679
Fehr E. and Gachter S. (2002) “Altruistic punishment in humans” Nature. 415 (6868) Jan 10:137-140
Ford, David (1982) “The development of buyer-seller relations in industrial markets” in H. Hakasson ed.
International Marketing and Purchasing of Goods – An Interaction Approach, New York: Wiley
Ford, David I. ed. (1997) Understanding Business Markets: Interaction, Relationships and Networks, London:
Dryden Press
12
Ford D and P J Rosson, 1982, The Relationships Between Export Manufacturers and Their Overseas
Distributors, in Czinkota M and G Tesar (eds), Export Management, Praegar, New York, pp. 257-275.
Ford D, A Lawson, J F Nicholls, 1982, Developing International Marketing Through Overseas Sales
Subsidiaries, in Czinkota M R, G Tesar (eds.), Export Management- An International Context, Praeger,
pp. 227-243.
Ford D and K Djeflat, 1983, Export Marketing of Industrial Products Buyer-Seller Relationships Between
Developed and Developing Countries, in Czinkota M R (ed), Export Promotion- The Public And
Private Sector Interaction, Praeger, pp. 262-284.
Ford J D and J W Slocum, 1977, Size, Technology, Environment and the Structure of Organizations, Academy
of Management Review, Oct, pp. 561-575
Florence, P. Sargeant (1933), The Logic of Industrial Organisation. London: Routledge and Kegan Paul.
Granovetter M, 1985, Economic Action and Social Structure: The Problem of Embeddedness, American
Journal of Sociology, 91, pp. 481-510.
Hakansson H, 1988, Formal and Informal Cooperation Strategies in International Industrial Networks, in
Contractor F, and P Lorange (eds), Cooperative Strategies in International Business, Lexington, Mass:
Lexington Books, pp. 369-80.
Hakansson H, 1987a, Corporate Technological Behavior: Cooperation and Networks, London, Routledge.
Hakansson H (ed.), 1987b, Industrial Technological Development- A Network Approach, Berkenham: Croom
Helm.
Hakansson, Hakan and Snehota, Ivan (1995), Developing Relationships in Business Networks.
Routledge.
London:
Hakansson, Hakan and Ivan Snehota (1998) The Burden of Relationships or Who’s Next, in Network Dynamics
in International Marketing, Peter Naude and Peter W. Turnbull eds., Pergamon, Oxford pp. 16-25.
Hallen L, J Johanson and N Seyed-Mohamed, 1991, Interfirm Adaptation in Business Relationships, Journal of
Marketing, 55, pp.29-37
Hallen L and F Weidersheim-Paul, 1979 Psychic Distance and Buyer Seller Interaction, Organization, Marknad
och Samhalle, 16
Hallen L and F Wiedersheim-Paul, 1981, Psychic Distance in International Marketing-An Interaction
Approach, Working Paper, Centre for International Business Studies, University of Uppsala, Sweden
Heckman, James J. (1979) “Sample Selection Bias as a Specification Error” Econometrica 47 153-161
Heide, Jan B. (1994) “Interorganizational Governance in Marketing Channels” Journal of Marketing, 58
(January), 71-85
Heide, Jan B. and John, George (1988) The role of dependence balancing in safeguarding Transaction -Specific
assets in conventional channels” Journal of Marketing 52 (January) 20-35
Heide, Jan B. and Milner, Anne S (1992) “The Shadow of the Future: Effects of Anticipated Interaction and
Frequency of Contact on Buyer Seller Cooperation” Academy of Management Journal 35 (2) 265-91
Hirsch S and Z Adar, 1974, Firm Size and Export Performance, World Development, 2 (7), pp. 41-46Klein B, R
G Crawford and A.A. Alchian, 1978, Vertical Integration, Appropriate Rents, and the Competitive
Contracting Process, Journal of Law and Economics, 21, October, pp. 297-326
13
Johanson J and LG Mattsson, 1986, International Marketing and Internationalization Processes - A network
approach, in Paliwoda S. and P.N. Turnbull (eds), Research in International Marketing, Croom Helm,
London.
Johanson J and L Mattsson, 1987, Interorganizational Relations in Industrial Systems- A
Network Approach Compared with the Transaction Cost Approach, International
Studies of Management and Organization, XVII(1).
Johanson, J. and Mattsson, Lars-Gunnar 1994: The Markets-as-Networks Tradition in
Sweden, Research Traditions in Marketing, in G. Laurent, G. Lillen and B. Prass eds.
Kluwer, Boston, MA. pp. 321-342
Johanson, J. and Mattsson, Lars-Gunnar (1988) “Internationalization in Industrial Systems A Network Approach” in N. Hood and J-E Vahlne eds. Strategies in Global
Competition. London: Croom Helm
Johanson J. and J.E. Vahlne, 1977, The Internationalization Process of the Firm: A Model of
Knowledge Development and Increasing Foreign Market Commitments, Journal of
International Business Studies, 8, Spring/Summer, pp. 305-322
Johanson J. and J.E. Vahlne, 1977 “The Mechanism of Internationalisation” International
Marketing Review 7 (4) 11-24
Johanson, J. and Vahlne, J-E (1992) “Management of Foreign Market Entry” Scandinavian
International Business Review, 1 (3) 9-27
Johanson, J. and Finn Wiedersheim-Paul (1975) “The Internationalisation of the Firm - Four
Swedish Case Studies” Journal of Management Studies 12 (3) 305-322
Kanuk L and C Berenson, 1975, Mail Surveys and Response Rates: A Literature Review,
Journal of Marketing Research, Vol.12, Nov., pp. 440-453
Kalwani, M.U. and N. Narayandas (1995), “Long Term Manufactuer-Supplier Relationships: Do They Pay Off
for Supplier Firms?.” Journal of Marketing, 59 1 1-16
Layton R A and D C Dunphy, 1970, Export Attitudes, Management Practices and Marketing
Skills, Australian Export Development Council.
Macaulay S, 1963, Non-Contractual Relations in Business: A Preliminary Study, American
Sociological Review, 28, pp. 57-66
McNeil I R, 1980, The New Social Contract, Yale, University Press, New Haven CT
MacNeil I, 1985, Relational Contract: What we do and and do not know, Wisconsion Law
Review, 3, pp. 483-526
Markusen, A. (1996) “Sticky Places in Slippery Space: a Typology of Industrial Districts”
Economic Geography 72 (3) 293-313
Martin, Xavier, Swaminthan, Anand, and Mitchell Will (1998) “Organizational Evolution in
the Interorginizational Environment: Incentives and Constraints on International
Expansion Strategy” Administrative Science Quarterly 43 (September) 566-601
Meyer, Klaus E. “Insitutions, Transaction Costs, and Entry Mode Choice in Eastern Europe” Journal of
International Business Studies 32 (2) 357-367
14
Miner, Anne S. (1992), "The Shadow of the Future: Effects of Anticipated Interaction and Frequency of Contact
on Buyer-Seller Cooperation," Academy of Management Journal, 35 (2), 265-91
Monteverde K and Teece D J, Appropriate Rents and Quasi-Vertical Integration, Journal of Law and Economic
25, pp.321-328
Munro H and P Beamish, 1987, Export Distribution and Performance, in P.J. Rosson and S. D. Reid, eds.,
Managing Export Market Entry and Expansion:
Concepts and Practice, New York,
Praeger.Nootebom, Bart (1999) Inter-Firm Alliances: Analysis and Design, London, Routledge
Pan, Yigang and Tse, David K. (2000) “The Hierarchical Model of Market Entry Modes” Journal of
International Business Studies 31 (4) 535-554
Parkhe, Arvind (1993) “Strategic Alliance Structuring: A Game Theoretic and Transaction
Cost Examination of Interfirm Cooperation” Academy of Management Journal, 36
(August), 794-829
Pedersen, Bent and Welch, Lawrence (2002) “Foreign Entry Mode combinations and
internationalization” Journal of Business Research 55 (February) 157-162
Petersen, Bent and Pedersen Torben (1997) “Twenty Years After - Support and Critique of the Uppsala
Internationalisation Model” in Ingmar Bjorkman and Mats Forsgren ed. The Nature of the International
Firm, Copenhagen: Copenhagen Business School Press, 117-134
Philips L W, 1981, Assessing Measurement Error in Key Informant Reports: A Methodological Note on
Organizational Analysis in Marketing, Journal of Marketing Research, 18, Nov., pp. 395-415
Porter, Michael E. Competitive Strategy: Techniques for Analyzing Industries and Competitors, New York, The
Free Press
Ragan, , V. Kasturi, E. Raymond Corey, and Frank Cespedes (1993) “Transaction Cost Theory; Inferences
From Clinical Field Research on Downstream Vertical Integration,” Organization Science, 4 (August),
454-77
Reid S.D., 1982, The Impact of Size on Export Behaviour in Small Firms, in Czinkota M. and
G. Tesar (eds), Export Management, Praegar Publishers, New York, pp. 18-38
Niraj, Gupta and Narasimhan (2001) “Customer profitability in a Supply Chain” Journal of Marketing, 65
(July), 1-16
Rindfleisch, Aric and Heide, Jan B (1997) “Transaction cost analysis: Past, present, and future applications”
Journal of Marketing 61 (October) 30-54
Rooks, Gerrit, Raub, Werner, Selten Robert and Tazelaar Frits (2000) “How Interfirm Cooperation Depends on Social Embededness: A Vignette Study” Acta Sociologica 43
(2) 123-137
Rosson, Philip (1984) “Success factors in Manufacturer-Overseas Distributor Relationships
in International Marketing” in Kaynak, Erdner ed. International Marketing
Management New York: Praeger 91-107
Rugman A.M., 1986, New Theories of the Multinational Enterprise: An Assessment of
Internalization Theory, Bulletin of Economic Research, 38(2), pp. 101-18
Shaver, J. Myles (1998) “Accounting for Endogeneity When Assessing Strategy
Performance: Does Entry Mode Choice Affect FDI Survival?” Management Science
44 (April) 571-585
Shelanski and Klein (1995) Empirical Research in Transaction Cost Economics: A Review
15
and Assessment” Journal of Law, Economics and Organisation 11(2) 335-61
Shapiro S P, 1987, The Social Control of Impersonal Trust, AGS, 93, pp. 623-58, .
Shaver, J. Myles “Accounting for Endogeneity When Assessing Strategy Performance: Does
Entry Mode Choice Affect FDI Survival” Management Science 44 (April) 571-585
Smits, Jereon (Undated) “Estimating the Heckman two-step procedure to control for selection
bias with SPSS.” http://home.planet.nl/~smits.jeroen/selbias/selbias.html
Stigler, George J. (1951), “The Division of Labor is Limited by the Extent of the Market.”
Journal of Political Economy, 95 (June), 185-193.
Teece D J, 1986, Transaction Cost Economics and the Multinational Enterprise, Journal of
Economic Behavior and Organization, 7, pp. 21-45
Thorelli, Hans and Cavusgil, S. Tamer eds. (1990) International Marketing Strategy 3rd
Edition New York, Pergamon Press, 129-139
Turnbull, Peter, David I. Ford, and Malcom Cunningham. 1996. Interaction, Relationships
and Networks in Business Markets: an evolving perspective. Journal of Business and
Industrial Marketing, 11(3/4): 44-62.
Von Hippel, Eric (1984) “ “Sticky Information” and the locus of problem solving: Implications for Innovation”
Management Science, 40 (4) 429-439
Welch and Wilkinson (in press) “Incorporating Idea Logics into Network Theory” Journal of Business to
Business Marketing
Wilkinson, Ian F. (2002) “A History of Channels and Network Thinking I Marketing in the 20 th Century”
Australasian Marketing Journal
Williamson Oliver E. (1975) Markets and Hierachies: Analysis and Antitrust Implications, NewYork: Free
Press.
Williamson Oliver E. (1981) “The Modern Corporation: Origins, Evolution, Attributes” Journal of Economic
Literature, 19 (December) 1537-1568
Williamson, Oliver E. (1983) “Credible Commitments: Using Hostages to Support Exchange” American
Econimic Review 73 (September) 83-87
Williamson, Oliver E. (1985) The Economic Institutions of Capitalism, New York: The Free Press
16
A Contingency Model of Entry Mode Performance
Other causes
Contingency
Variable
Fit
Entry Mode
Organisation
+
Performance
Feedback
source: adapted from Donaldson (2001) p12
17
Table 1
Classification of countries by Cultural and Geographic Distance from Australia
a) Cultural distance
Very close (code = 1): South Africa, New Zealand, UK, Italy, France, Spain, USA, Canada,
Norway,Federal Republic of Germany, Sweden.
Medium (code = 2): Papua New Guinea, Fiji, South Pacific Islands, Hongkong, Taiwan, Japan,
Indonesia, Malaysia, Philippines, Singapore, Thailand, South Korea, People's Republic of
China.
Very distant (code = 3): Brunei, Burma, Cambodia, Laos, Macao, Vietnam, North Korea,
Afghanistan, Pakistan, Mongolia, India, Sri Lanka, Bangladesh, Nepal, Bhutan, Noumea,
Solomon Islands, Tonga, New Caledonia.
b) Geographic Distance
Very close (less than 5000 kms, code =1): New Zealand, Papua New Guinea, Fiji, New Caledonia,
Noumea, Solomon Islands, Tonga.
Medium (5000 to 10,000kms, code = 2): South Pacific Islands, Hongkong, Taiwan, Japan, Indonesia,
Malaysia, Philippines, Singapore, Thailand, Vietnam, North Korea, South Korea, China.
Very distant (greater than 10,000kms, code = 3): UK, Burma, Pakistan, Afghanistan, India, Sri
Lanka, Turkey, Kuwait, Iran, Iraq, Israel, Lebanon, Greece, Eastern Block, African Nations,
USA, Canada, Norway, Sweden, Austria, Belgium, Mexico, Switzerland.
18
Table 2
Factor Analysis of Product Characteristics
(Maximum Likelihood with Varimax Rotation Chi Sq = 7.5 n.s.)
Product Attributes
Product is unique, differentiated or represents
advanced technology
The production process is exclusive to our firm
Product has significant advantages over
competition
The Company R&D level exceeds industry average
Product requires extensive training to operate and
use
Product requires considerable after sales support
Principal Product is the same product as is sold
domestically
Principal Product is a standardized product
applicable to most foreign markets without any
modification
Alpha for underlined items
Factor 1
.76
Factor 2
.19
Factor 3
.01
.67
.57
-.01
.14
-.03
.08
.56
.10
.24
.76
-.08
-.09
.21
-.11
.66
-.10
.09
.80
.07
-.01
.41
.76
..69
.46
19
Table 3
Perceived Risk Scale
Items
Exporting should only be considered when opportunities in
Australia are completely exhausted
There is too much risk involved in exporting for my
organisation to be engaged in it
The quality of my company’s products could never be good
enough to sell in overseas markets
Exporting is too different from marketing in Australia to
enable my organization to succeed
Alpha
Loading
.71
.86
.75
.83
0.87
20
Table 6
Two Stage Model of Export Entry Mode: Stepwise Logit Regression
mode
Stage 1
Stage2A
Int’mediary or
Domestic or
Vert Int.
Foreign
Int’mediary
Domestic or Foreign Intermediary = 0 (n)
150
37
Direct Exports or Foreign Sales Branch =1 (n)
133
113
Stage2B
Direct versus
Foreign Sales
Branch
78
55
Variable
Intercept
-.095
1.526***
.441
Product Asset Specificity
Hitech
Service
Foreign Market Size and Duration of involvement
Sales to Market
Years exporting to Market
Uncertainty
Cultural Distance
Geographic Distance
Perceived risk
Firm Scale, Scope and Internationalization
Firm Size
Percentage of Exports
No. of countries exported to
No. of Products Exported
Type of Industry
Consumer
Industrial
Interaction Effects
a) Uncertainty*Asset Specificity
Cultural distance*Hitech
Cultural distance*Service
Geographic distance* Hitech
Geographic distance* Service
Perceived risk *hitech
Perceived risk *service
b) Uncertainty *Duration
Cultural distance*years in market
Cultural distance*years in exporting
Geographic distance* years in market
Geographic distance* years in exporting
Perceived risk * years in market
Perceived risk * years in exporting e
-.184***
-.608***
.002**
.002**
.442**
.103**
.161**
.077**
-.093*
22.66 (4)***
.077
64%
3.369(1)*
.022
75%
b) Market involvement* Duration
Foreign Market Sales*years in market
Foreign Market Sales *years in exporting
Chi Square (df)
Cox and Snell R Square
Percent Correctly classified
*=P<.1 **=P<.05 ***=P<0.01
46.42***
.295
74.4%
21