The Differing Foreign Entry Mode Choices for Sales and Production Subsidiaries of Multinational Corporations in the Manufacturing Industry

Foreign market entry mode research has been a popular area of study. However, a clear agreement between the usage of conventional constructs and their impact on a firm’s entry mode choice has not yet been found. This paper focuses on how, depending on the type of subsidiary that is established, multinational corporations (MNCs) in the manufacturing industry use different foreign market entry strategies. Previous research either treated types of subsidiaries synonymously or investigated them separately. However, due to the changing competitive landscape and disaggregation of value chain activities into separate subsidiaries, I find it necessary to compare how these entry mode choices differ depending on the activity each subsidiary is responsible for. My analysis finds that MNCs in the manufacturing industry are more likely to use joint ventures rather than wholly owned modes of entry for their production subsidiaries in comparison to their sales subsidiaries. I further explore how the international experience of the MNC strengthens this effect. This research utilizes a sample of 201 listed Korean manufacturing firms and 833 foreign market entry mode choices into 49 countries.


Introduction
Research on entry modes into foreign markets has been carried out extensively due to the lasting decisions and significant impacts these choices have on consequent firm performance and firm sustainability [1][2][3][4]. Although entry mode research has been carried out in depth through a variety of perspectives and theories such as transaction cost theory [5,6], the eclectic paradigm [7,8] and cultural distance [9], scholars have maintained the need for more entry mode research [10]. This is because the results acquired from previous studies provide no explicit agreement regarding the effect certain variables have on entry mode decisions into foreign markets [11]. In addition to this, scholars in the field of international business have stated that the question of how firms should enter foreign markets is one of the most central questions to be answered for internationalization [12].
When trying to answer the question of how firms should enter foreign markets, scholars seem to be broadly separated into two streams of research [3]. While one group of scholars has allocated contracts, joint ventures and wholly owned subsidiaries along a continuum of increasing control, commitment and risk [13,14], another group of scholars have found entry mode choices to be based on how firms want to compensate their contributors of business or input providers [15][16][17]. This latter stream of research suggests that firms tend to establish joint ventures with a local partner when the required resource for foreign entry is difficult to purchase in the market. Therefore, when these resources are difficult to transact, the resource holder is paid ex-post from the profit accrued by the venture. This study builds on this latter stream of research to examine how the different characteristics and input needs between sales and production subsidiaries within multinational corporations (MNCs) in the manufacturing sector can impact their foreign entry mode decisions.
Although previous literature has accounted for the differences in these types of subsidiaries and how these subsidiaries differ in their location choices or sequence of foreign entry when internationalizing [18], studies comparing the differences between these subsidiaries in their foreign entry mode decisions have been less apparent [19,20]. The Uppsala model has explained that firms are more likely to establish sales subsidiaries before production subsidiaries when internationalizing because they aim to steadily increase their commitment or investments [18,21,22]. As the international business landscape has changed and developed, so have the roles and competencies of MNC subsidiaries [23]. Rugman, Verbeke & Yuan (2011) [23] categorized Porter's (1985) [24] well-known classification of value chain activities into four broader groups which included innovation, production, sales and administration. In concurrence with new global trends, scholars have suggested that instead of viewing value chain activities in collective terms we should account for the disparities that are apparent across them [23,25]. These new trends such as technological advancements and globalization have led to the breakdown and separation of value chain activities, whereby some MNCs have begun to carry out each activity in separate subsidiaries or locations [23,[26][27][28]. Therefore, although activities such as production and sales were once viewed to be closely related to each other, they are now able to function and be established independently from each other [23]. In sequence with this view, I believe a more detailed investigation of the foreign entry mode decisions pursued by MNCs for their different types of subsidiaries will provide meaningful insights within the field of international business.
While differences regarding foreign entry mode decisions between the manufacturing and service industry have been investigated [29], this research paper places emphasis on the differences that are inherent between the types of subsidiaries that exist within the manufacturing industry. Specifically, it focuses on how MNCs in the manufacturing industry employs different foreign market entry strategies depending on whether they are establishing a sales subsidiary or a production subsidiary. Previous research on foreign entry modes seems to have investigated this issue separately, as they seemed to have focused on one activity in the value chain at a time [19]. For instance, early research from entry mode scholars concentrated on the manufacturing industry and their production activities [30], while marketing scholars focused on the comparison between sales subsidiaries and local distributors [31]. This paper aims to investigate the distinct characteristics and roles sales and production subsidiaries hold within manufacturing MNCs to predict foreign entry mode choices as a result of their different input requirements. In particular, I argue that this difference arises from the fact that sales subsidiaries require more human inputs while production subsidiaries require more capital inputs. I also examine how the international experience of the manufacturing MNCs impacts these differences.
Overall, the research setting utilizes listed MNCs in the manufacturing industry from Korea and their foreign market entry mode choices with regard to the type of subsidiary that is being established. I focus in particular on sales and production subsidiaries. I believe using the context of Korean manufacturing MNCs is ideal for several reasons. First, in the past few decades, Korean manufacturing companies have transformed and grown at a rapid pace to become strong competitors at a global level [32]. Some examples of highly reputable Korean MNCs in the manufacturing industry that have achieved a high status worldwide include Samsung Electronics, Hyundai Motors, LG electronics and Kia Motors [32,33]. I also believe that Korean MNCs in the manufacturing industry provides a sufficient level of variety regarding the number of countries where foreign direct investments were pursued. The final sample is comprised of 833 FDIs into 49 countries by 201 listed MNCs in the manufacturing industry from Korea, of which 425 production subsidiaries were established and 408 sales subsidiaries were established. In addition to this, the MNCs in this sample had an even allocation of production and sales subsidiaries, whereby the MNCs had an average of two production subsidiaries and two sales subsidiaries each. This provided an ideal context to test my theories. Although a comprehensive list of individual mode structures for internationalization has been found to exist, I followed a high number of studies which concentrated on the choice between wholly owned subsidiaries (WOSs) and joint ventures (JVs) [3]. Furthermore, a review by Shen, Puig & Paul, (2017) on entry mode research has found that most of the research on entry mode decisions focused on the choices of firms from developed countries [34]. Therefore, as research utilizing the Korean context is limited and Korean MNCs in the manufacturing industry have become significant players in the global context, I believe further investigation and focus on the Korean context is needed.
My theory and supportive empirical findings on the differing foreign entry mode strategies pursued by manufacturing MNCs for their production and sales subsidiaries provide important contributions. My arguments on the relationship between different subsidiary types and foreign entry mode choices display how their different functions can influence manufacturing MNCs to carry out differing foreign entry mode decisions when internationalizing. This is due to the differing emphasis they place on certain inputs when choosing to operate abroad. The detailed investigation into production and sales subsidiaries extends previous literature on foreign modes of entry which primarily focused on one activity in the value chain at a time [19]. The empirical analysis and results indicate that manufacturing MNCs are more likely to use joint ventures rather than wholly owned modes of entry for their production subsidiaries in comparison to their sales subsidiaries. Furthermore, the results indicate that the international experience of MNCs moderates the relationship between subsidiary types and foreign market entry mode decisions. Therefore, this study extends the significance of examining not just the type of industry an MNC is in but the types of subsidiaries that exist within an industry and MNC, as this can lead to distinct input needs which can further contribute to their resulting foreign entry mode decisions.
This study also provides important implications for managers and policymakers. FDIs have been found to positively affect the sustainable development of host countries [35]. Therefore, by understanding the different modes of entry firms prefer when conducting FDIs, policymakers will be able to create better regulations that could help the economic sustainability of their countries.
In particular, the government should also acknowledge the different types of roles or functions foreign subsidiaries will carry out when established. In addition to this, FDIs are seen as a form of sustainable growth for firms. This is because it has been viewed that firms expand to international markets to improve their performance and thereby improve their sustainability [36]. Therefore, learning how firms internationalize by examining the modes of entry firms employ could also help managers with important decisions regarding FDIs which could further aid firm growth.

Manufacturing MNCs' Entry Mode Choice by Subsidiary Type
A wide number of studies on foreign market entry mode have focused on both the manufacturing industry [37,38] and service industry [39]. While these studies have provided relative support that the primary theories explaining entry mode choices can be applied to both manufacturing and service industries [29], differences regarding foreign entry mode choices between the two industries have been noted [6]. These disparities stem from the fact that firms in different industries place a differing amount of importance to certain variables [29]. For instance, firms in the manufacturing industry can be considered to be more capital intensive, since they need to invest in plants and equipment. On the other hand, firms in the service industry can be seen to require less financial investment when entering into foreign markets as they are more people-intensive and require lower amounts of investment in plants and equipment [29].
The notion of different subsidiary types can be fairly similar to the different industries in that sales subsidiaries can also be seen to be more dependent on human capital and production subsidiaries on capital investments. Internationalization studies have accounted for the differences in types of subsidiaries as they have found some firms to follow the strategy of gradually increasing their levels of risk and commitment to foreign countries, making them more likely to establish sales subsidiaries before production subsidiaries [18,40]. Furthermore, it has been found that subsidiaries in MNCs hold unique roles and classifications, with different resources being allocated to the various types of subsidiaries [41,42]. Although MNC subsidiaries were once seen to carry out several value chain activities bundled together [42], recent trends have allowed for the specialization and separation of activities [23,26,43,44]. The specialization and separation of activities in subsidiaries have also allowed for individual activities to be conducted separately [27,28]. While different scholars have categorized value chain activities based on various criteria, I focus on two of the four activity categories set forth by Rugman, Verbeke & Yuan (2011) [23]; production activities which include inbound logistics such as operations and sales activities which include outbound logistics such as marketing, sales and services.
Therefore, as value chain activities have been disaggregated [23,25,45], subsidiaries in charge of certain activities have been seen to hold unique and different roles from each other. In this regard, MNCs could prefer different foreign entry modes for subsidiaries in charge of particular value chain activities due to the unique roles and functions they are responsible for. Consequently, as sales subsidiaries are in charge of more downstream activities [23], they would require higher levels of local knowledge [46]. The knowledge sales subsidiaries require to carry out key activities such as marketing, sales and services can be seen to be more location-bound in comparison to the competencies that are needed by production subsidiaries [46,47]. As this knowledge is seen to be location bound, the advantages and competencies need to be developed at the host country and it is difficult to transfer these skills to new markets [48]. Furthermore, firms have been known to use market research companies to acquire data on customers to conduct sales [49]. Also, to obtain customers, it is important for firms to recognize who their customers are and what their needs are which is why local knowledge is essential [49]. Therefore, I believe that much of the knowledge and competencies sales subsidiaries will need to succeed can be obtained through the hiring of local talent.
On the other hand, production subsidiaries require high fixed costs due to the high level of investment that is needed to set up production processes and to develop knowledge to carry out their roles [46]. As production subsidiaries are in charge of manufacturing products, they may require certain assets such as land and facilities which can be challenging to obtain. In addition to this, firms can be hesitant in investing considerable amounts of resources when entering a foreign market due to the risk of failure and high levels of uncertainty [50]. Therefore, it can be argued that manufacturing MNCs may perceive higher levels of risk and uncertainty for their production subsidiaries due to heavy capital investment that is needed to set up production. Furthermore, I believe firms would want to share the burden of this investment with local partners due to the high levels of risk and uncertainty involved. Through this perspective, it can also be argued that sales subsidiaries can more easily obtain the necessary resources they require to be competent in comparison to production subsidiaries. This is because I believe it might be more difficult to procure the lands and facilities that production subsidiaries need in contrast to hiring local talent which is what I believe sales subsidiaries may require. Therefore, as joint ventures are typically preferred when assets are difficult to obtain and wholly owned subsidiaries are preferred when assets are easier to obtain [51], I predict that manufacturing MNCs will be more prone to using joint ventures for their production subsidiaries in comparison to sales subsidiaries. The arguments above lead to the following hypothesis regarding foreign market entry mode choices for manufacturing MNCs with regard to their sales subsidiaries and production subsidiaries.

Hypothesis 1 (H1).
Manufacturing MNCs are more likely to use joint ventures rather than wholly owned modes of entry for production subsidiaries in comparison to sales subsidiaries.

Moderating Role of International Experience
Previous literature and empirical studies have found international experience to affect a firm's foreign market entry mode decisions. These studies have argued that levels of experience are important for internationalization and foreign market entry mode decisions because lack of knowledge may Sustainability 2019, 11, 4089 5 of 18 contribute to higher levels of uncertainty and costs [52]. Studies have found firms to evade higher equity modes of entry such as wholly owned subsidiaries into countries whereby they faced higher levels of uncertainty [38]. This is because the uncertainty and risks involved could lead to higher levels of cost associated with entering the country [53]. Therefore, as experience could help firms acquire knowledge and reduce levels of uncertainty, this could also reduce the levels of reliance firms would hold on their local partners, which would lead to higher equity modes of entry [54,55].
However, I believe international experience will moderate the likelihood of using joint ventures differently depending on whether the subsidiary is a production subsidiary or sales subsidiary. I argued previously that much of the knowledge and competencies sales subsidiaries would need to succeed can be acquired through the hiring of local talent as they require high amounts of local knowledge. I also argue that higher levels of international experience will further reduce the levels of uncertainty or lack of knowledge these subsidiaries may have and therefore diminish their need for local partners [56]. Consequently, I argue that international experience will strengthen the likelihood of manufacturing MNCs using wholly owned subsidiary modes of entry for sales subsidiaries.
I also previously argued that manufacturing MNCs might prefer using joint venture forms of entry for their production subsidiary. I further argue that international experience may strengthen this effect as higher levels of experience will also increase the levels of knowledge the firm acquires and could make it easier to monitor or coordinate with their local partners [56]. Therefore, the issues that firms may have had because of the cultural differences that were inherent with their local partners [6] may be less of a problem. In addition to this, I also argued that the characteristics of a production subsidiary might require them to acquire certain assets such as land and facilities which could be challenging to obtain in foreign markets. However, international experience may also enable manufacturing MNCs to find local partners with the resources they need more efficiently. Therefore, I argue that as the level of international experience of the MNCs increases, this will strengthen the likelihood of them using joint ventures rather than wholly owned modes of entry for production subsidiaries in comparison to sales subsidiaries.

Hypothesis 2 (H2).
International experience will strengthen the likelihood that manufacturing MNCs will use joint ventures rather than wholly owned modes of entry for production subsidiaries in comparison to sales subsidiaries.

Sample and Data
My sample consists of publicly listed Korean MNCs in the manufacturing industry and their foreign subsidiaries. Data on the foreign subsidiaries was obtained from a survey conducted by the Korea Trade Promotion Corporation (KOTRA) in the year 2016. The survey which is titled the "Overseas Korean Business Directory" provides comprehensive information on the overseas subsidiaries of Korean MNCs including information on subsidiary types, entry modes, host country locations and the establishment years of foreign operations.
From the original data, I was able to identify 1395 foreign subsidiaries of publicly listed Korean MNCs that were in the form of FDIs. These Korean MNCs were listed on the Korea Composite Stock Price Index (KOSPI). For the purpose of this study, I excluded subsidiaries that belonged to parent firms in the financial or service sector. Therefore, only subsidiaries of parent firms that belonged to the manufacturing sector were kept. Parent firms from the manufacturing industry were identified according to the Korean Standard Industry Classification code. At the end of this process, only 912 foreign subsidiaries remained. I also dropped subsidiaries that were missing information (such as entry years, subsidiary types and foreign entry modes) as well as subsidiaries that belonged to parent firms that did not close their accounts in December. The final sample totaled up to 833 foreign subsidiaries distributed over 49 countries worldwide. Financial data of the parent firms were retrieved from KISVALUE, which is a widely used data source for studying publicly listed Korean firms. In addition to this, economic data of the host countries the subsidiaries were established in was obtained from The World Bank website.
The final sample consists of 201 manufacturing MNCs from Korea with an average of four foreign subsidiaries each. Furthermore, the sample consists of 425 production affiliates and 408 sales affiliates. A detailed summary of the statistics of the manufacturing MNCs at the parent firm level can be found in Table 1 while the distribution of the countries and types of foreign subsidiaries can be found in Table 2.
The statistics of the parent firms in Table 1 displayed an even allocation of production and sales subsidiaries, whereby the MNCs had an average of two production subsidiaries and two sales subsidiaries each. This even distribution of sales and production subsidiaries provided a good context to test my theory on subsidiary types and entry mode decisions. Furthermore, as shown in Table 2, there was a high variation in the number of host countries FDIs were carried out into.
Additional analysis on the country distribution reveals that 292 FDIs were made in advanced economies (35% of total FDIs) (I followed IMF's country groupings based on level of economic development [57]). Among these, 20.5% of total FDIs were comprised of production subsidiaries while 79.5% were sales subsidiaries. In contrast, 541 FDIs were made in middle income and developing economies (65% of total FDIs). Among these, 67.5% were comprised of production subsidiaries while 32.5% were sales subsidiaries. According to this sample, more production subsidiaries were located in middle income and developing economies presumably to retain lower production input costs.  (1) and wholly owned subsidiaries as (0). I used the entry mode information given in the Overseas Korean Business Directory to code the dependent variable.

Independent Variables
Subsidiary type. To test the first hypothesis, I further coded my main independent variable as a binary variable. I categorized production subsidiaries as (1) and sales subsidiaries as (0). Service subsidiaries have also been included as sales subsidiaries as their function is an auxiliary one to sales subsidiaries in general.
International experience. To test the second hypothesis, I counted the number of years each parent MNC had operated their business internationally. This allowed me to measure the international experience of the parent firms. By following the approach of [45], the variable was measured by subtracting the year when a parent firm's first FDI had occurred from 2016 which is the observation year.

Control Variables
Various conventional control variables at the subsidiary level, parent firm level and country level that could impact foreign entry mode choices were also included.
The subsidiary level controls included subsidiary age. I follow the notion of Hennart (1991) [55], whereby the established mode of entry may change as the subsidiaries age and a joint venture could dissolve or be acquired by a partner. Since the data set is cross-sectional, there is a possibility that some of the subsidiaries may have initially been a joint venture but fully acquired by Korean parents. Subsidiary age was included in the investigation to correct for such possible changes in entry modes. Subsidiary age was measured as the number of years that had passed since the establishment of the subsidiaries in their host countries.
The parent firm level controls were obtained and coded from the financial database KIS-VALUE which is the Korean version of the U.S. COMPUSTAT financial database. I first controlled for firm age as a proxy for the general experience of parent firms in their business.
Parent firm's R&D intensity and advertising intensity were included in order to control for the level of proprietary assets of the parent firms. Prior literature on entry mode suggests that firms with higher levels of proprietary assets are likely to choose high control modes of entry to avoid local partner's opportunistic behaviors such as free-riding or dissemination of technology [38]. R&D intensity and advertisement intensity were measured by calculating their ratio over total sales of the firms (for example, advertisement spending of 2016/total sales of 2016).
In addition, firm size (total number of employees in natural log) was measured to control for the parent firm's size effects. Past literature suggests that larger firms prefer higher control modes of entry since they have abundant resources that can be utilized in new markets and they can avoid risks that can emerge from making partnerships with local partners [39]. Lastly, several host country level controls such as GDP growth (3-year average), GDP (in natural log), geographic distance (in natural log) in addition to cultural distance were included.
GDP growth was included in the models to control for market potentials [58]. For instance, in fast-growing markets, high growth in demand would provide firms with high potentials for growth. Firms would more likely enter with full modes of entry to appropriate full returns from investments rather than share them with local partners. In a similar vein, host country GDP has been included in the models to control for host country market size [59].
Geographic distance and cultural distance are also controlled in the models to control for the perceived uncertainty of investing firms. Higher levels of physical distance from the home market and large cultural differences may provide uncertainty for managers. Therefore, larger geographic distances may generate information asymmetry between local partners and firms and as a result, make it more challenging for firms to monitor their local partners [60]. Moreover, with higher cultural differences, Korean firms may find difficult to understand their local partners' actions in the host country due to unfamiliar cultural environments [61]. Hence, they may prefer wholly owned mode of entry when there are higher levels of geographic and cultural distances. Geographic distance was measured as the physical distance between the capital of Korea (Seoul) and the capital cities of the host countries. The log of the geographic distance values was calculated due to its dispersed distribution. For the measure of cultural distance, Kogut & Singh (1998)'s [62] standardized measure was used, which is an aggregation of the standardized distances of each of Hofstede's cultural dimensions. A detailed description of the variables can be found in Table 3.

Data Analysis
To test the hypotheses, I utilized the logistic regression model to analyze the cross-sectional data. I applied the following model to examine the effect of the independent variables on the probability of a firm choosing joint venture mode of entry: where Y is the dependent variable, x is the vector of independent variables of the logit regression, b is the vector of coefficients of the logit regression and a is the intercept of logit regression. I followed the majority of empirical studies that used cross-sectional designs due to the difficulty in acquiring data that would allow for longitudinal studies [3,63,64]. Table 4 shows the descriptive statistics and correlation analysis for the main variables. The correlation analysis suggests that the variables were within reasonable levels of correlation with the highest correlation being −0.66 (between GDP growth rate and geographic distance). In addition to this variance inflation factors (VIFs) were also checked for and they were all found to have values that were less than 5. This suggested that multicollinearity was not an issue in the empirical models. The mean of the dependent variable entry mode was 0.127 which shows that most of the foreign direct investments in the sample were conducted through wholly owned subsidiaries. However, this was not a problem when running the logistics regression since the models had a good fit with the data in the sample. The main independent variable subsidiary type was evenly distributed with a mean of 0.510. Lastly, the mean of the moderating variable international experience variable was 28.18 which show that the average international experience of the MNCs in the sample was 28 years.

Analysis of Manufacturing MNCs' Entry Mode Choice by Subsidiary Type
The models in my analysis presented in Table 5 provide information about the direction of the influence my independent variables have on a Korean MNC's choice of entry mode. Positive coefficients indicate that the independent variable increases the probability of a firm choosing a joint venture mode as its entry strategy. On the other hand, negative coefficients indicate that the dependent variable decreases the probability of a firm choosing a joint venture mode of entry. Model 1 shows the effects of the control variables and how they impact entry mode decisions while Model 4 contains the full model excluding the interaction term. Models 2 and 5 display the results which pertain to my research objectives. Model 2 tests hypothesis 1 and Model 5 presents the result for hypothesis 2 whereby the interaction between subsidiary type and parent firm's international experience is formally tested.
The models overall have a high explanatory power with valid chi-square values. They also show a high classification rate over-predicting wholly owned modes of entry. This is partly due to the large portion of wholly owned modes of entry in comparison to joint ventures modes of entry in the sample. However, all the models that present the essential results of my research objectives, excluding models 3 and 4, satisfy the Hosmer-Lemeshow test. This proves that my models have a good fit with the data in the sample.
The results strongly supported the theoretical perspectives of this study. Model 1 tested for all the control variables at the subsidiary, parent and host country level and how they impacted entry mode decisions. Some of the control variables such as parent firm size, geographic distance and cultural distance had a significant impact on entry mode decisions. Larger parent firms were associated with joint venture modes of entry and firms were less likely to use joint venture modes of entry into countries that were geographically further away. Both of these variables were statistically significant (p < 0.01). Cultural distance was also negatively related to a firm's choice of joint venture mode of entry (p < 0.01). This suggests that when firms are entering culturally distant countries, they tend to establish wholly owned entities and avoid partnering with local firms. All three variables were significant in all models in this study. Lastly, at a 10% significance level, host country GDP levels were negatively related to joint venture occurrences. Firms were less likely to establish joint ventures in countries with large market sizes.
Consistent with Hypothesis 1, manufacturing MNCs were more likely to use joint ventures rather than wholly owned forms of entry for their production subsidiaries in comparison to their sales subsidiaries. The results for the main hypothesis were strongly supported and statistically significant with a positive coefficient as shown in Table 5 Model 2 (p < 0.01). Therefore, hypothesis 1 was strongly supported.
Consistent with Hypothesis 2, international experience positively moderated the relationship between production subsidiaries and their likelihood of using joint ventures as opposed to wholly owned modes of entry. This result was also strongly supported and statistically significant (p < 0.01). The data analysis was interesting as it showed that international experience did not directly affect the entry mode decisions of subsidiaries. However, the results became statistically significant when the international experience of the MNC at the parent firm level was interacted with the type of subsidiary in the manufacturing MNC to impact the foreign entry mode choice that was pursued.
To obtain a better understanding of the findings, I plotted an interaction graph for the second hypothesis. To compare the different probabilities of joint venture selections for each subsidiary type, I calculated the adjusted predictions (predictive margins) for sales and production subsidiaries at different international experience year values with other variables set to their means. These calculations were then plotted in Figure 1 to show the different probability conjectures for each subsidiary type. Figure 1 illustrates that there is a moderating effect between the international experiences at the parent firm level and the relationship between subsidiary types and foreign entry modes choices. The figure illustrates that an increase in international experience by year units increases the probability for joint venture selection for production subsidiaries whereas the probability decreases for sales subsidiaries. Figure 1 shows a clear difference in preferences for entry mode choices between production subsidiaries and sales subsidiaries. Figure 1 also clearly shows how their preference is affected by a parent firm's international experiences.

Discussion
The current study investigated how MNCs in the manufacturing industry used differing foreign entry mode strategies for production and sale subsidiaries and I came across some interesting findings. The results strongly supported the first hypothesis which suggested that MNCs in the manufacturing industry were more likely to use joint ventures rather than wholly owned modes of entry for production subsidiaries in comparison to sales subsidiaries. I theorized that these different modes of entry stemmed from the distinct needs of the two subsidiary types. I argued that sales subsidiaries were more dependent on local knowledge which could be provided by the hiring of local employees, whereas production subsidiaries were more in need of local partners to assist them with their capital costs and retaining production inputs. Furthermore, the results provided support for the second hypothesis whereby as the international experience of the MNCs increased, this strengthened the likelihood of them using joint ventures rather than wholly owned modes of entry for their production subsidiaries in comparison to sales subsidiaries. I argued that as parent firms gained more international experience, they would learn to monitor or coordinate their partners better and as a consequence the cultural differences that could have contributed to the difficulty of (3.

Discussion
The current study investigated how MNCs in the manufacturing industry used differing foreign entry mode strategies for production and sale subsidiaries and I came across some interesting findings. The results strongly supported the first hypothesis which suggested that MNCs in the manufacturing industry were more likely to use joint ventures rather than wholly owned modes of entry for production subsidiaries in comparison to sales subsidiaries. I theorized that these different modes of entry stemmed from the distinct needs of the two subsidiary types. I argued that sales subsidiaries were more dependent on local knowledge which could be provided by the hiring of local employees, whereas production subsidiaries were more in need of local partners to assist them with their capital costs and retaining production inputs. Furthermore, the results provided support for the second hypothesis whereby as the international experience of the MNCs increased, this strengthened the likelihood of them using joint ventures rather than wholly owned modes of entry for their production subsidiaries in comparison to sales subsidiaries. I argued that as parent firms gained more international experience, they would learn to monitor or coordinate their partners better and as a consequence the cultural differences that could have contributed to the difficulty of operating joint ventures for their production subsidiaries may no longer be an issue. In contrast, I argued that more international experience would reduce the level of uncertainty for sales subsidiaries and strengthen the likelihood for them using wholly owned modes of entry.
These findings have several implications and contributions to previous literature. I believe the comparison between production and sales subsidiaries helps bridge the gap between theory and practice. The detailed investigation into the comparison between production and sales subsidiaries was an extension of previous foreign entry mode studies which generally focused on one activity in the value chain at a time, whether it was production, R&D or sales [19]. Although MNC subsidiaries were once seen to conduct a collection of activities, many of them are now seen to specialize or focus on one activity [23,[26][27][28]. This study contributes to this changing competitive landscape whereby value chain activities have been disaggregated [18]. This has led to the need for further investigation into the different types of subsidiaries, allowing me to contribute to the developing internationalization literature whereby subsidiaries in MNCs are becoming the unit of analysis [65]. In addition to this, the majority of entry mode research has been conducted to examine the manufacturing or production issues of internationalization [19]. Therefore, I believe this study investigates how there could be alternative factors or circumstances which could impact existing arguments. For instance, joint ventures have been found to help firms obtain local knowledge through their local partners [66] and having a local partner has been found to help firms overcome their liability of foreignness which includes the different costs incurred by a firm when operating businesses abroad [67]. However, this study suggests that the means of acquiring local knowledge may differ depending on the value chain functions foreign subsidiaries are responsible for. Unlike previous studies that argued joint ventures were an efficient way to obtain local knowledge through local partners, I argued that the local knowledge required by sales subsidiaries could be obtained through the hiring of local employees. Therefore, taking into account the different types of subsidiaries that exist, wholly owned modes of entry could perhaps be a more efficient mode of entry for sales subsidiaries. It was also interesting to note how international experience moderated the likelihood of using joint ventures differently depending on the subsidiary type. For Korean MNCs in the manufacturing industry, it became more likely for them to use joint ventures for production subsidiaries as they gained more international experience. This was in contrast to previous studies that found firms to use higher equity modes of entry as experience led to lower levels of uncertainty [54,55]. However, Korean MNCs followed this logic for their sales subsidiaries when international experience increased due to the reduced need for local partners.
It is important for managers and practitioners to consider these factors as past research has found that the performance of subsidiaries or the sustainability of subsidiaries can be affected by whether or not they have used the correct form of entry when internationalizing [1,68,69]. Therefore, when managers are deciding to move or establish their operations abroad, they should first consider what type of subsidiary they are planning to have or what function their subsidiary will mostly be in charge of. The difference between whether the subsidiary is a production subsidiary or sales subsidiary could impact the efficiency of their entry mode decision. Moreover, methods of obtaining the knowledge required to succeed may differ depending on whether a sales subsidiary or production subsidiary is being established. I believe this investigation and results using empirical data from Korean MNCs in the manufacturing industry will also be helpful for managers and practitioners as Korean manufacturing companies have become globally competitive in the past few decades. Examining their modes of entry could enlighten managers from other countries on the different strategies their firms should enact depending on the type of subsidiary they want to establish abroad. Furthermore, the context of Korean MNCs provided an ideal sample and context to compare the difference between subsidiary types due to the even allocation of production and sales subsidiaries.
In addition to this, policymakers should take these factors into account when creating new regulations for their country's economic sustainability as FDIs have been found to have positive effects on economic growth [35]. In order to create reasonable regulations to generate an attractive environment for businesses and FDI inflows, policymakers need to understand how foreign companies will want to establish their operations. In addition to concentrating on the types of industries or characteristics of firms, the government should also acknowledge the different roles or functions these subsidiaries will carry out when established. For instance, as this study predicts MNCs may use wholly owned modes of entry for their sales subsidiaries because they can obtain local knowledge through the hiring of local employees, governments could consider creating attractive environments for sales subsidiaries as their inward FDIs and their establishments could lead to more jobs for the local population. This is an important factor to consider as past literature has found that FDI has positive effects on growth when interacted with human capital [70]. Furthermore, as production subsidiaries are predicted to use joint venture forms of entry and seek local partners, this will also be a good avenue for the economic sustainability of a country. This is because production subsidiaries may transfer their technological know-how or capital to their local partners. Therefore, governments should also create favorable regulations which would attract production subsidiaries to help their local firms.

Conclusions
While there has been a plethora of entry mode studies [3,63,64], the present study proposed subsidiary types to affect the entry mode decisions of manufacturing MNCs in foreign countries. Although studies on entry mode choices have investigated the entry mode decisions of manufacturing firms, service firms and the comparison of these two industries [3,29], less attention has been paid to a comparison between the differences posed by subsidiaries within the same industry [19]. This study examines the difference between sales and production subsidiaries and how they relied on different inputs for operations. Although I argued that it was more critical for sales subsidiaries to rely on local talent, I also argued that production subsidiaries were required to rely more on local partners due to the higher levels of investment they may need. I investigated how depending on these different types of subsidiaries there could be alternative factors or circumstances which could impact existing arguments on entry mode research.
Past research has suggested that firms with localization objectives were more likely to form joint ventures with local partners to gain access to local culture and knowledge. According to past logic, sales subsidiaries which require higher levels of local knowledge should be entering foreign markets in the form of joint ventures. On the other hand, production subsidiaries which are situated upstream in the value chain need to be better coordinated and controlled with their headquarters and therefore should more likely be established through wholly owned modes of entry. My investigation provided some opposing views and results to this theory, suggesting that sales subsidiaries preferred wholly owned modes of entry while production subsidiaries preferred joint venture modes. These interesting results can be interpreted with the fact that local knowledge for sales subsidiaries could be obtained through the hiring of local talent, mitigating the need for local partners. However, setting up production subsidiaries through the recruitment of local talent may be more complicated as the level of risk and investment required is much higher. In this aspect, production subsidiaries may rely more on local partners to acquire institutional knowledge which would be difficult to acquire by merely employing local talent. Issues that require local institutional knowledge such as labor issues and host country government relationships could also be responded to proficiently with a local partner's assistance. Local partners could also help investing firms adapt to the host country's institutional environments [1,71]. My results are in support of this argument and provide evidence that production subsidiaries are more likely to seek local partners.
Therefore, the differences in entry modes that have previously been recorded as industry effects [29] could be due to the different characteristics of the subsidiaries comprised in each industry. In my additional analysis I found that although MNCs in the manufacturing industry held a substantial variation between production and sales subsidiaries, MNCs in the service industry consisted predominantly of sales subsidiaries. This is not surprising given that service industries are in charge of intangible products whereby there is no physical production of their products [72].
Although I believe this paper contributes to the entry mode stream of research, it has several limitations. First, this study only investigates the sales and production subsidiaries of MNCs in the manufacturing industry. However, research and development subsidiaries could be another interesting type of subsidiary to examine as innovation is another important value chain activity. Second, this study pertains to MNCs in the manufacturing industry from Korea. Therefore, only entry mode decisions by one country were examined. Although there is much to learn from Korean MNCs in the manufacturing industry as many of them have become strong performers in the global market over the past few decades, it could be interesting to investigate whether entry mode choices differed depending on whether parent firms were from emerging countries or developed countries and the regions they were internationalizing to. It would be interesting to investigate if further differences in entry modes existed depending on the region firms were internationalizing to as the supplementary analysis on the country distribution of FDIs revealed 35% of total investments were made in advanced economies while 65% of total investments were made in middle income or developing economies. In addition to this, an extension of this study could be to investigate the survival, sustainability or performance of these subsidiaries depending on whether they used the more efficient mode of entry. Furthermore, although this study only investigated two groups which were sales and production subsidiaries, these groups could also be segmented even further. For example, some production subsidiaries may be responsible for massive production whereas other production subsidiaries may be responsible for niche productions which would require more knowledge or skills [20].
Therefore, investigating how the entry mode decisions depended on the types of subsidiaries in MNCs may open a gateway for subsequent research. The empirical analysis provided strong support that within MNCs in the manufacturing industry, production subsidiaries were found to prefer joint ventures in comparison to sales subsidiaries. Furthermore, although this study showed how international experience strengthened this effect, alternative and more important influences could exist. Policymakers and managers should take into account how MNCs use different modes of entry for their production and sales subsidiaries, as this could affect and help improve the sustainability of both firms and economies.