Does the Foreign Market Entry Mode Choice Affect Export Performance? The Case of the Spanish Hotel Industry

Knowing the factors that affect the export performance has become one of the most recent research lines in the literature relative to business internationalisation. In this sense, several works based on the meta-analysis technique have come up in the two last decades in order to recapitulate the set of studies which have analysed such a topic. Taking these reviews as a starting point, we may establish that there is not a clear relationship between the entry mode and export performance variables. With a study on a sample of 122 Spanish hotels all over the world, we have the intention of analysing the existence and sense of the relationship between both variables. As a conclusion, we may state that the market entry mode has a positive and significant influence on export performance.


Introduction
An important issue to be considered when a company is going to internationalize is the entry mode through which the fi rm decides to serve foreign markets (Root 1987). The set of options to penetrate other markets is pretty broad and varied, with several ways of action depending on the parameters that are given more importance to. In all of these ways of action, the level of resources endowment and the degree of risk assumed are moving in the same direction. Nevertheless, this is not the case of the degree of control, which is increased as the resources endowment goes up at the same time that fl exibility decreases.
The specialist literature offers a set of works which suggest that the proper election of the entry mode would have a positive impact on the performance (Chen, Hu 2002;Brouthers 2002;Aulahk, Kotabe 1997). Thus, those companies that, according to their own characteristics, do not choose correctly the entry mode to foreign countries will become less effi cient and, depending on market forces, could be expelled by the competition (Brouthers 2002). Therefore, the relationship between entry mode and performance variables could be analysed in the following way: fi rstly, it would be interesting to study if the entry mode decision has been correctly made by the surveyed fi rms according to the theoretical prescription; and secondly, it would also be attractive to compare the performance level between those companies that chose correctly the entry mode and those fi rms that did not. This twofold objective will allow us to analyse in an integrative way the decision concerning the entry mode and its consequences on the performance level (Sanchez, Pla 2003).
There are studies such as the one carried out by Brouthers (2002) that try to shed light on this topic. Brouthers (2002), taking the transaction cost theory as a starting point, reveals differences in export performance among the different market entry modes. More recent studies such as the one developed by Ekeledo and Sivakumar (2004), based on the Resource Based View, or the one made by Rasheed (2005), also reveal the differences in export performance due to the entry mode election. Hence, the election of an entry mode is possibly related to export performance.
Bearing in mind the considerations above mentioned, this work has the intention of deepening in the analysis of the factors which affect the export performance, but paying a special attention to the relationship with the entry mode variable. To do so, this work is structured as follows: after this introduction, we are going to expose the set of theories more relevant to both variables (entry mode and export performance); secondly, we will carry out an analytic review on the literature about export performance. In this sense, we will employ the meta-analysis technique in order to fi nd out, in recent years, which factors are affecting the export performance variable and determine its relationship with the entry mode variable; next, we will describe the research methodology employed in our study, and fi nally, we will show the results obtained with our analysis and describe the main conclusions we reach as a result of our work.

Relevant theories on entry mode and export performance
In this section, we are going to make reference to two different theoretical fi elds: one related to the entry mode topic, and the other one linked to the export performance analysis. Both fi elds offer complementary explanations to better understand how the process of internationalisation is carried out and how the best entry mode may be selected when entering foreign markets.

The entry mode topic
Within the wide scope of theories concerning business internationalisation, two specifi c theories conclusively explain and analyse the entry mode topic. These theories are the Internalization Theory (among others : Brouthers 2002;Shrader 2001;Liesch, Knight 1999;Caves 1996;Buckley, Casson 1988;Anderson, Gatigon 1986) and the Resource Based View of the fi rm (among others: Kalnins, Chung 2004;Madhok 1997;Kogut, Zander 1993) The entry mode in foreign markets is a classical issue within the specialist literature on international businesses. Greatly infl uenced by the transaction cost theory, traditional studies consider each particular entry as a 'transaction' (Anderson, Gatigon 1986), and the most important aspect has to do with trusting the external options of the market (such as exportation) or, on the contrary, the internal options developed by the company (for instance, the direct investment) (Buckley, Casson 1988). On the other hand, the Resource Based View of the fi rm highlights the level of analysis of the transaction of the company suggesting that the decision of the particular entry cannot be considered in isolation. It must be considered with respect to the set of strategic positions of the company (Hill et al. 1990).
Thus, the transaction cost theory analyses which entry mode is able to minimise the transaction costs linked to the exploitation of a competitive advantage in a foreign market. Such exploitation may be contracted between external agents (modes of low control level) or may be internalised within the fi rm (modes of high control level). The company will opt for those modes which imply a higher control level when the markets where the fi rm is able to negotiate its competitive advantages do not work properly and, consequently, it is necessary to incur high transaction costs in order to safeguard a misuse of these advantages if there were possible exchanges with other agents, such as the case of licences or joint ventures (Buckley 1995).
In the case of the hotel trade, the higher level of control on the external operation permits to mitigate the tendency towards the opportunism on the part of the hotels in a twofold sense: a) property rights offer a greater potential to establish a richer rewarding system; and b) the organisational culture shared by the chain of hotels and its hotels in property provides with a set of norms and values more aligned with the interests of the chain (Brown, Dev 2000).
From the Resource Based View of the fi rm, when those assets or processes liable to be internationally exploded are based on a tacit 1 know-how it will be predominant the investment in full-property branches, since such assets are very diffi cult to be transferred, sometimes, without the transference of the human team or even the own organisation. Bearing this argument in mind, it may be justifi ed the effi ciency of the company to transfer such knowledge with a lower cost to its branches than to an alien organisation. On the other hand, those knowledge turns into platforms which facilitate the growth and expansion outside, due to its novelty and diffi culty to be imitated (Ekeledo, Sivakumar 2004).

The export performance topic
With respect to the literature on export performance, the Resource Based View of the fi rm is of utmost importance to analyse the attitude of fi rms towards the international activity. In fact, this theory is especially relevant when we try to understand the mode of expansion chosen by the fi rm in its international adventure (Ekeledo, Sivakumar 2004). Obviously, this theory is not the only one to try to explain which entry mode is the most suitable one for every market (since there are many different factors that explain this decision), and it is not exclusive to try to understand why companies decide to become international.
Therefore, we fi nd support in this theory in order to explain the study of both the export performance and the market entry mode. In this sense, the work by Dhanaraj and Beamish (2003) shows a rigorous model, based on this theory, which relates the internationalisation strategy and export performance. The Resource Based View of the fi rm is focused, in its essence, on how a superior performance may be achieved, in comparison to other competitors in the same market, thanks to the exploitation and acquisition of unique resources by the fi rm. This perspective is valid due to the fact that the specialist literature develops a rich theoretical framework that contributes to generate and test models of internationalisation.

Determinant factors of export performance
There is not unanimity in the specialist literature concerning the factors, and their relative importance, which affect the performance of the foreign activity ). This fact is due to the great fragmentation and shortage of general conclusions in this research arena (Li et al. 2008;Eusebio et al. 2007;La et al. 2005;Aaby, Slater 1989;Cavusgil, Zou 1994). From this viewpoint, it turns out to be extremely diffi cult to derive the business behaviour patterns that guide the execution of foreign trade operations, as well as the construction of a normative framework which supports the decision-making process. Table 1 shows in a very summarised way the contributions made on the topic of determinant factors of export performance linked to the export activity. Taking this framework as a starting point, the identifi cation of these determinants in our sample becomes one of the goals of our research. As Leonidou et al. (2002) state, the attention in this kind of research has been paid to fi ve groups of variables that infl uence export performance. These groups are the following: • Managerial variables (staff, experience, attitudes, behaviour, features of who makes the decisions related to foreign activities); • Organisational variables (characteristics, operations, resources, objectives to be achieved with foreign activities); • Environmental variables (factors, whether macro or micro-environmental, that affect the organisation, and that infl uence both the domestic and the international market); • Market-goal variables (identifi cation, selection and segmentation of international markets); and • Marketing-mix variables (price, placement, publicity and product that the company commercialises overseas).  Leonidou et al. 2002) These fi ve groups of variables and their relationship with export performance have been conceptualised in some studies in the specialist literature (Cavusgil, Zou 1994;Gemünden 1985), and are employed by posterior works in order to frame the different factors affecting export performance (Li et al. 2008;Eusebio et al. 2007;La et al. 2005). The analysis carried out by Leonidou et al. (2002) contributes to determine a model where each set of variables affect export performance. As a result, a fi rst group of variables including managerial, organisational and environmental variables serve as a background to the strategy and indirectly infl uence export performance. A second group encompass those variables related to the marketing strategy (market-goal and marketing-mix variables), and are directly related to export performance. And a third group puts together the measures of the export performance variable (both economic and non-economic ones).
Hence, according to these studies it may be stated that, explicitly, the foreign market entry mode variable is not revealed as a determinant factor of export performance. Ne vertheless, bearing in mind the set of factors affecting the entry mode in foreign markets, which may be implicitly represented in the fi rst four groups of variables infl uencing export performance (Leonidou et al. 2002), we may deduce that there is a relationship (direct or indirect) between entry mode and export performance variables.
Likewise, in this work we have tried to analyse the review of specialist literature through the development of a descriptive meta-analysis 2 in order to check how the studies on determinant factors of export performance have been evolving. Though the list of determinant factors is pretty long, we are not able to derive conclusions for all the relationships since there is not enough number of studies. This fact prevents us from inferring the existence of a signifi cant relationship between many of the factors and the export performance variable. However, in the case of some factors we have found enough works to be able to establish the existence of a statistically signifi cant relationship between these factors and the export performance variable, as the size of the effect is big enough to derive such conclusions.
We assume as a premise that we will only be able to generalise the relationship between two variables when, under different contexts or circumstances, we are able to achieve the same results (Bilkey 1987). Likewise, we may group the different factors depending on the fact that they have had a positive or negative infl uence on export performance. In addition, we may also cluster the factors in a third group, that is, those factors which have not a signifi cant relationship with export performance (because the number of studies is not great enough, or because the conclusions derived from them are not explanatory at all). As a result, we may fi nd studies where a signifi cant relationship (positive or negative) do exist and works where do not. Table 2 illustrates the kind of effect of each determinant factor on export performance, as well as the strength of its relationship.
As shown by our theoretical results, there is not a clear relationship between entry mode and export performance variables. In our opinion, two explanations may be underlying this fact: fi rstly, the number of studies that have been developed is not great enough so that the conclusions derived from them may be born in mind; and secondly, in the results obtained by the works analysed we fi nd that there are, in similar proportions, conclusions about positive and signifi cant relationship, negative and signifi cant relationship, and non-signifi cant relationship. Precisely, this latter fact has led us to try to deepen in the study on the relationship between market entry mode and export performance. Hence, as a summary of this analytic study of the determinant factors of export performance we may conclude that there is not a clear relationship between this variable and the market entry mode variable. This fact is on the basis of our main research goal, as we intend, through the development of an exhaustive analysis, to shed a bit of light on this issue, that is, the link between the results obtained as a result of the development of activities in foreign markets and the market entry mode choice. The lack of clear conclusions in this topic does not allow us to establish the relationship between both variables in any sense. So, we enunciate the relationship between them in all the possible ways: H 1A : There is a positive and signifi cant relationship between the market entry mode choice and export performance H 1B : There is a negative and signifi cant relationship between the market entry mode choice and export performance H 1C : There is not signifi cant relationship between the market entry mode choice and export performance

Research methodology
We have chosen the Spanish hotel industry to carry out the empirical work of this study due to the relative importance that this sector has got for the Spanish economy. Also, we have taken the model proposed by Contractor and Kundu (1998) as a starting point. This model contributes to determine which the most suitable market entry mode is according to a set of determinant factors.
On a sample formed by 122 Spanish hotels all over the world, we have identifi ed which of them selected, as the market entry mode, the choice proposed by the model and which did opt for modes different from the ones prescribed by the model. Afterwards, we have compared the performance obtained by the hotels between those fi rms that followed the prescriptions of the model and those companies that did not.
Concerning the information gathering stage, we have turned to primary information sources through the employment of the questionnaire. We set in touch with the 622 establishments that all the Spanish chains of hotels had got around the world in 2006. As a result of the process, we received 127 replies, 5 of which were ruled out because they were incomplete. Finally, 122 questionnaires were considered valid.
The entry mode variable has been measured in a polynomial way. That is, we have chosen the different market entry modes proposed by Contractor and Kundu (1998) and have divided them into four groups or modes: fully Owned, Partially Owned (joint venture), Management Service Contract, and Franchises. These four modes have been categorised according to the amount of resources (from great to small) committed to the development of foreign operations on the part of the chain of hotels.
On the other hand and as a result of the meta-analysis carried out, we established that the best way to measure the export performance variable was through a mixture of fi nancial and non-fi nancial indicators. Specifi c studies (e.g. Sousa 2004) give support to this way of measuring export performance. The measure we employed was obtained subjectively thanks to the assessment of top managers. According to Geringer and Hebert (1991), subjective performance measures are preferred when objective fi nancial measures are not available. And some works that have been reviewed (Brouthers et al. 2003;Woodcock et al. 1994) reveal the enormous diffi culty of obtaining objective measures of export performance in the case of foreign branches and, as a consequence, suggest the employment of subjective measures.
Subjective measures are able to measure the variable from a fi nancial viewpoint, a non-fi nancial perspective or a mixture of both (Dess, Robinson 1984). Both types of measurement, fi nancial (Shrader 2001;Pan, Chi 1999; among others) and non-fi nancial (Brouthers et al. 2003), have been analysed in previous works on the market entry mode variable. One of the limitations of the fi nancial measures has to do with the fact of comparing them among subsidiaries that were set up in different moments of time (Glaister, Buckley 1998;Geringer, Herbert 1991). In order to avoid this problem we have employed as a measure the performance average since the establishment was set up. On the other hand, non-fi nancial measures offer relevant information relative to strategic and competitive goals (Anderson et al. 2002). But, there is a limited consensus in the specialist literature regarding the existence of valid strategic and competitive measures of export performance. Consequently, prior studies on this topic have opted for combining both, fi nancial and non-fi nancial measures.
For example, the work by Brouthers (2002) employed three fi nancial measures (sales, profi tability and sales growth) and four non-fi nancial measures (market share, marketing, reputation and market access), all of them adopted from prior studies (Brouthers et al. 1999(Brouthers et al. , 2000(Brouthers et al. and 2002Geringer, Herbert 1991;Dess, Robinson 1984). All these measures were subjectively assessed and were measured in a 0 to 10 Likert scale (0 = any expectation is covered; 10 = totally covers all the expectations). The factorial analysis confi rmed the validity of both dimensions of export performance: the fi nancial one ( = 0.82) and the non-fi nancial one ( = 0.87). Additionally, we have considered the work by Brouthers and Nakos (2005) where measures of expectation and satisfaction are analysed compared to results achieved by competitors in the same industry. Bearing in mind all these considerations, we have opted for measuring the export performance variable taking as a basis the indicators mentioned above, paying special attention to the indicators employed by Brouthers (2002). This decision is coherent with a research on the hotel industry (Morey, Dittman 2003), where the measurement of the performance is based on measures of satisfaction of top managers. We confi rmed the reliability of the selected indicators ( = 0.78).
To sum up, we have established the way of measuring the export performance variable stemming from the analysis carried out about the empirical literature concerning export performance. In the set of the analysed works, we have made out that export performance variable has been measured through: objective measures (fi nancial and non-fi nancial) and subjective measures (perception about results). Thus, we have been forced to employ mixed measurements in order to determine the performance level of foreign operations. (for non-fi nancial indicators) were the most similar works to our idea of research for the work presented here. Finally, we have selected for our study four of the measures proposed by Brouthers (2002) (three of the fi nancial ones, concretely, sales, profi tability and sales growth; and one of the non-fi nancial measures, specifi cally, market share), as well as another additional measure taken from the work by Min- Han and Min-Kim (2003), that is, the degree of satisfaction with the global results. Though all these items and their measurement scales have already been validated in their corresponding studies, they have had to be validated again since they have been adapted to our work. Let's remember that for our study the  value is equal to 0.78.

Results
We have employed the ANOVA 3 technique in order to analyse the relationship between market entry mode choice and export performance variables. The fi rst thing we did was to test the normality of the variables to be employed. All the indicators that formed the independent variable export performance (let's remember that export performance is a mixed variable) were scored on a 0 to 10 Likert scale (eleven positions). Later, we proceeded to act the same as Min and Min (2003), both in the election and the treatment, with respect to the indicators that formed the export performance variable.
Through the observation of asymmetry and kurtosis coeffi cients we checked that the export performance variable did not tend to normality, and after the application of the Kolgomorov-Smirnov test we reached the conclusion that this variable did not present a normal distribution. In order to solve this problem, we tried to reach an adjustment of the original variable towards a normal probability distribution. After several attempts where we tried different powers which would provide us with better results to achieve the goal of normality, we established that the power to three was the most suitable one. Later, the observation of asymmetry and kurtosis coeffi cients as well as the Kolgomorov-Smirnov test revealed that we had achieved that the export performance variable followed a normal probability distribution.
After checking that both variables were fulfi lling the requirements of homocedasticity (through the Levenne statistic) and normality, we proceeded to contrast the relationship between the market entry mode choice and export performance variables. Therefore, we had the intention of testing the existence (or not) of signifi cant differences in the results of those Spanish hotels that had entered a foreign market following the recommendations of the original model proposed by Contractor and Kundu (1998) compared to those Spanish hotels that had chosen an entry mode which did not follow such recommendations. Table 3 shows the results of the ANOVA test for these two groups of fi rms.
As it can be seen in Table 3, after the homogeneity test of variances between market entry mode choice and export performance variables we may state that signifi cant differences do exist between the results obtained by both groups of Spanish hotels. Concretely, a better level of performance was achieved by those establishments which chose the market entry mode suggested by the original model of Contractor and Kundu (1998). Hence, the relationship between both variables could be established as it is represented in fi gure 1, being confi rmed, as a result, the hypothesis 1A stated before.
Consequently, those internationalised Spanish chain of hotels that chose, for each hotel competing in a foreign market, a market entry mode following the proposal of the model by Contractor and Kundu (1998) were observed to have not only signifi cant differences but also a greater level of performance, compared to those hotels that opted for an entry mode different from the one proposed by such a model when carrying out their internationalisation process.

Conclusions
The work presented here has revealed the strength of the transaction cost theory and the Resource Based View of the fi rm to explain the selection of the access to international markets.
Through an analytical study we have been able to identify the relationship among the different determinant factors of export performance. Among these factors, the market entry mode choice did not evidence a clear relationship with the export performance variable. Also, as a result of this analytical theoretical study we have found some works where a signifi cant and positive relationship between entry mode and export performance did exist, some studies that revealed a signifi cant and negative relationship between them and some research that did not show any signifi cant relationship between both variables. This ambiguity in the effect of the market entry mode choice on export performance has been one of the primary drivers in our interest for analysing in depth this relationship.
Together with this lack of clear effect in the relationship between both variables, the absence of a number great enough of works where the joint study of both variables has been carried out has led us to establish that no relevant conclusions concerning the role played by the entry mode as a determinant factor of export performance may be derived. By and large, the relationship between entry mode and export performance variables has been analysed separately. In other words, some studies reveal that the choice of a particular entry mode entails a greater probability of success in international operations. Nevertheless, these works do not make clear whether these choices have been made according to the theoretical prescriptions.
Precisely, this is the main contribution of our work. We have analysed the importance of achieving the fi t between the theoretical proposals and the entry mode decision when a fi rm has to enter a foreign market. In the paper presented here we have revealed that a positive and signifi cant relationship does exist between the market entry mode choice and export performance. Taking the predictive variables of the original model as a starting point, we have separated those Spanish hotels that followed the guides of a theoretical model on entering foreign markets from those hotels that made the entry mode choice differently. We have compared the results of both groups in our sample of 122 Spanish hotels and, as a consequence, we have identifi ed signifi cant differences between them both. Explicitly, by comparing business performance indicators we have been able to give support to the original model and its convenience for the Spanish hotel industry. Our work presents two clearly defi ned groups of implications. On one hand, regarding the theoretical implications we have got to broaden the time horizon in the study of the determinant factors of export performance. Additionally, we have deepened in the analysis of a particular relationship, namely, the link between the market entry mode choice and export performance. On the other hand, the practical implications of this work are clearly applicable to the Spanish hotel industry. Firms in this particular industry will fi nd extremely useful to follow the model proposed by Contractor and Kundu (1998) with regard to the modal choice for each property, since the results of our study evidence the existence of positive and signifi cant differences in export performance between those hotels that do follow its prescriptions and those hotels that do not. Thus, for the top managers of this kind of establishments the model we have presented in this paper turns out to be a good tool to guide the internationalisation process of Spanish chains of hotels.
Despite the results we have obtained, this study is not exempt from limitations. First of all and concerning the theoretical analysis, the meta-analysis we have developed is merely descriptive through the employment of the vote counting technique. It would be more suitable to carry out a quantitative meta-analysis. This fact would contribute to the application of more complex statistical techniques and generate results which permit some kind of generalisation. Secondly, we have observed very few non-signifi cant results in the reviewed works. One possible explanation to this fact has to do with the publication bias. In other words, works with signifi cant results are more likely to be accepted for publication than studies with non-signifi cant results. Hence, a meta-analysis only containing published works is possible that offers biased results in the sense of an overestimation of the interest effect. And thirdly, another limitation could be derived from the measurement of export performance. Though we have based this measurement on a mixed measure (as a consequence of the results of the meta-analysis application), this fact is not guaranteeing that the employed indicators were the best ones to reach the goal we had established for our research. Therefore, we assume that another different measure of export performance could modify the results of our work.
Finally, future research lines could be oriented to keep on analysing which factors do carry on being important in order to improve the export performance measurements. Likewise, it would also be motivating to compare the results obtained in our research to the results of another type of industry, as well as employing different measures of evaluation of the results.