Performance of European Joint Ventures in Latin America, Asia and Eastern Europe

Transcription

1 Performance of European Joint Ventures in Latin America, Asia and Eastern Europe Paper submitted at the Third Iberoamerican Academy of Management São Paulo, December 2003 International Management Track Paper Session Pierre-Xavier MESCHI (lead author) Professeur Euromed Marseille School of Management Domaine de Luminy BP Marseille Cedex 9 France pxmeschi@univ-aix.fr Franck BRULHART Maître de Conférences Faculté de Sciences Economiques et de Gestion Université de la Méditerranée 14 Avenue Jules Ferry Aix-en-Provence Cedex France f.brulhart@univ-aix.fr Frédéric PREVOT Professeur Assistant Euromed Marseille School of Management Domaine de Luminy BP Marseille Cedex 9 France prevot@esc-marseille.fr

2 Abstract The aim of this paper is to examine the internationalization joint ventures formed in emerging countries. More precisely, this consists in studying their survival rate and to relate it to variations in the relevant country risk of the emerging countries concerned. we will detail the method used for gathering and analyzing data and the characteristics of a sample of 139 internationalization joint ventures that were formed during year 1996 by foreign partners from the European Union and local partners belonging to different emerging countries (mainly Latin American, Asian and eastern European countries). The survival analysis was applied to a sample of 139 joint ventures, together with a log-logistic analysis of the impact of different country risk variables on the hazard rate of these joint ventures. Time-related variations were observed in their termination rate (i.e., inverted double U variation) and their hazard rate (i.e., curvilinear distribution with honeymoon effect). Concerning the impact of country risk, the results obtained provided only a partial confirmation of this impact: the relation between country risk and the joint venture survival is confirmed empirically, but only when the emerging country concerned undergoes deterioration of its risk as time progresses. In conclusion, the relation between these two variables exists but it is a discontinuous one.

3 Introduction Why do companies get involved in joint ventures? Considering the large amount of work done on identifying the motives of the partners in joint ventures, this question seems almost trivial. The topic of the partners motives corresponds to a research field that has been explored in every possible detail. Categorizations of joint ventures have even been drawn up, based on the criterion of the partners motives. Most of these categorizations traditionally distinguish two major types of joint venture that each correspond to a specific motive: firstly, additive joint ventures (Garrette & Dussauge, 1995) and scale joint ventures (Hennart, 1988) whose aim is to attain economies of scope and scale; and secondly, complementary (Garrette & Dussauge, 1995) or link (Hennart, 1988) joint ventures that involve partners who are looking to combine distinct but complementary competencies within a joint project. These typologies were then refined and complemented by adding new types of joint ventures based on less general and more specific motives. In particular, this was the case with joint ventures of the Trojan Horse type, in which the partners engage in a learning race in order to acquire some of its ally s (or allies ) competencies as quickly as possible (Hamel et al., 1989; Hamel, 1991), and also restructuring or transition joint ventures (Nanda & Williamson, 1995) that enable one of the partners to discreetly get rid of a line of business that it considers marginal in relation to its core business. This refined detailed analysis of the partners motives must not make us forget that the primary historic function of joint ventures is to enter emerging countries that are considered difficult to get into, for geopolitical, regulation-related or cultural reasons. By being initially careful to consider that a joint venture is formed whenever two or more sponsors bring given assets to an independent legal entity and are paid for some or all of their contribution from the profits earned by that entity (Hennart, 1988: 362), it is useful to recall that the first joint venture experiments were carried out in the early years of the Soviet Union. It was against a background of the end of a

4 civil war and serious economic crisis that Lenin authorized the formation of mixed capital ventures (with mixes of State/private and foreign/soviet capital), as of November Thanks to this new form organization, certain American oil companies Chevron Oil, Socony and Standard Oil of New Jersey, which resulted from the dismantling of the Standard Oil Trust (i.e. the famous baby Standard Oils ) were able to exploit oilfields in the Baku area and Caspian Sea area, in association with Soviet State companies. In general, these joint ventures involving internationalization or expansion into emerging countries are formed and organized as follows: the foreign partner, from a developed country, provides the joint venture with upstream resources such as funding and production technology (Connolly, 1984; Pan, 1996), and the local partner provides downstream resources such as familiarity with local markets, access to the distribution network and personnel (Inkpen & Beamish, 1997; Kale & Anand, 2001), and knowledge of local regulations and preferential access to the public powers (Kale & Anand, 2001). The aim of this paper is to examine these internationalization joint ventures formed in emerging countries. More precisely, this consists in studying their survival rate and to compare it with that of other types of joint ventures. Starting from the idea that this type of joint venture, more than any other, is affected by variations in the economic and political conditions of emerging countries, we also strove to relate their survival rate to variations in the relevant country risk of the emerging countries concerned. This paper is presented in the following parts: in a first section, we will present an analysis of existing research literature concerning the duration (longevity) and the survival rate of different types of joint ventures, and their most significant explanatory factors. The question of the impact of the country risk on these variables will also be presented in this part. In a second part, we will detail the method used for gathering and analyzing data and the characteristics of a sample of 139 internationalization joint ventures that were formed during year 1996 by foreign partners from the

5 European Union and local partners belonging to different emerging countries. The third part will combine the results of the survival analysis applied to the 139 joint ventures of the sample, together with a log-logistic analysis of the impact of different country risk variables on the hazard rate of these joint ventures. In the last part, the results will be discussed by highlighting the survival rate of internationalization joint ventures and the most significant explanatory variables. This last part will also discuss the specific character of this type of joint venture. 1. Joint Venture Survival and Duration: Definition and Impact of Country risk Past research into joint ventures has frequently addressed the question of the performance of joint ventures. The measurements of this performance are diverse (measurements of economic and financial factors, stock market factors, satisfaction, duration, etc.) and they adopt different viewpoints (i.e., analysis of the performance from the viewpoint of the partners, or the viewpoint of the joint venture itself). But, far from opposing each other, these studies are part of a more general approach of triangulation of the performance of joint ventures. Duration and survival rate are some of these measurements that are used to evaluate performance, and more precisely the success of joint ventures. These two variables were the first measurements of the success of joint ventures, and they are still predominant in research literature. Franko s study in 1971 initiated a whole school of research into the performance of joint ventures, based on survival analysis methodologies (Elandt-Johnson & Johnson, 1980; Cox & Oakes, 1983; Freeman et al., 1983) and considering that there is a strong, positive correlation between the performance, the duration, and the survival of a joint venture (Pearce, 1997: 203). This study is a standard reference work for studies of the survival of joint ventures because it was the first to formalize the different situations that may cause termination of the alliance. On the basis of a sample of 1,100 joint ventures formed between 1961 and 1968 by American companies

6 investing abroad, Franko (1971) measured a 28.5% termination rate, and therefore a 71.5% survival rate, on the basis of the identification of three types of modification in the capital structure of these joint ventures during the analyzed period: 1) firstly, the American partner increases its share in the joint venture to more than 95%, thereby transforming it into subsidiary. This situation amounts to a form of termination of the alliance that may be described as acquisition or internalization of the joint venture. 2) Then the American partner increases its share in the joint venture, but to a level below 95%. 3) Finally, the American partner transfers all its share to its partner, or all the partners agree to dissolve the joint venture. This situation corresponds to another form of termination of the alliance, which is pulling out of the joint venture. Of these three types of modification of the capital structure, one situation (i.e., 2) above) is often left out of survival analyses because it refers to a form of the joint venture instability, but it is difficult to call it a real termination of the alliance. This is because, in this situation, as opposed to the other situations (i.e., 1) and 3) above), the joint venture continues to exist essentially in the same form as before its capital modification. An analysis of existing research studies on the survival of joint ventures shows that some will completely adopt (Blodgett, 1992; Barkema & Vermeulen, 1997) or slightly adapt (Kogut, 1988, 1989 et 1991; Park & Russo, 1996; Park & Ungson, 1997; Leung, 1997; Hennart et al., 1998; Shi, 1998; Delios & Beamish, 2001) the variables of the joint venture termination presented in Franko s study (1971). The main adaptation concerns the scope of variables that measure the joint venture termination. Therefore certain studies have limited measurement of the termination by only studying forms of acquisition and/or pulling out of joint ventures. This is the case of studies by Kogut (1988 & 1991), Park & Russo (1996) and Shi (1998) in which the failure of joint ventures is measured by the number of dissolutions and acquisitions. Leung (1997), Hennart et al. (1998) and Delios & Beamish (2001) chose to measure termination by pulling out (sell-offs

7 and dissolutions) of joint ventures. Kogut (1989) and Park & Ungson (1997) limited measurement of the termination to only dissolutions. Table 1 presents a summary of survival studies that explicitly identified duration, survival rate and sometimes the form of variation in the annual hazard rate of joint ventures. On these different points, results in Table 1 are found to be in agreement. Firstly, it is seen that around 15% of joint ventures end in their first two years of existence, and around 50% in the six years following their creation. The analysis of the annual termination rate defined as the ratio of (a) the number of joint ventures that ended during the studied year, and (b) the total number of joint ventures that survived reveals two periods during which there is a higher increase than normal in this rate. The first high-risk period is during the third year of existence, and the second during the sixth or seventh year. Independently of the reasons that can cause dissolution, acquisition or pulling out of a joint venture, it is interesting to note that the average survival of the joint venture seems to be especially threatened during these two periods. Each of these periods corresponds to a critical phase of destabilization and questioning of the joint venture. Table 1 also shows that conclusions on the duration, the survival and the high-risk periods of joint ventures are virtually the same, irrespective of the nature domestic or international of joint ventures. Consequently, these conclusions may be extended to internationalization joint ventures, even if certain results in Table 1 concerned joint ventures of which one essential function was to enable foreign companies to enter the United States. In order to identify variations in the duration and the survival rate of joint ventures, some studies developed and tested explanatory models including a varying number of variables. The variables most frequently tested by these studies include: equity ownership (Blodgett, 1992; Park & Russo, 1996; Park & Ungson, 1997), the domestic or international nature of the joint venture (Kogut, 1988; Park & Russo, 1996; Park & Ungson, 1997; Shi, 1998), the cultural gap (Park &

8 Ungson, 1997; Barkema & Vermeulen, 1997), the existence of previous relations between partners (Kogut, 1989 et 1991; Park & Ungson, 1997), the joint venture size (Kogut, 1988; Park & Russo, 1996; Park & Ungson, 1997; Shi, 1998), and the partners previous experience of the joint venture s investment mode i.e., their mode experience (Park & Russo, 1996; Delios & Beamish, 2001) and of the country or countries in question i.e., their country experience (Barkema & Vermeulen, 1997; Hennart et al., 1998; Delios & Beamish, 2001). To our knowledge, the impact of country risk on the joint venture survival was only tested in the study by Barkema & Vermeulen (1997). On the basis of a sample of 828 joint ventures formed between 1966 and 1994 by 25 Dutch multinational companies that invested in 72 different countries, this study showed that the country risk did not significantly affect the survival of internationalization joint ventures. However, we find it difficult to formulate a definitive conclusion on the impact of country risk on the basis of a single study focused on internationalization joint ventures formed by Dutch companies and to apply this conclusion generally to all internationalization joint ventures. Furthermore, other factors lead us to further examine the question of the impact of country risk on the survival of joint ventures and to envisage new tests to confirm such results. Studies using another measurement of the performance of internationalization joint ventures i.e., stock market reactions to announcements of the formation of these joint ventures identified a significant negative stock market impact of country risk (for the two countries concerned, China and Russia) for American partners (Cheng & MacDonald, 1998) and French partners (Meschi & Hubler, 2003).

9 Study Size and Characteristic of the sample 1. Kogut (1988) 148 domestic and international joint ventures including at least one American partner 2. Kogut ( domestic and international & 1991) joint ventures including at least one American partner 3. Park & Russo (1996) 4. Park & Ungson (1997) 204 domestic and international joint ventures including at least one American partner 186 domestic and international joint ventures including at least one American partner 5. Leung (1997) 214 international joint ventures including at least one American partner 6. Shi (1998) 121 domestic and international joint ventures Table 1 Summary of studies on the joint venture duration and survival Market Period Studied* Survival rate Variation in the Annual Termination Rate ** USA Inverted double U variation with a first peak at 3 years (9.3%) and a second peak after 6 years (20.9%) USA % after 2 years Inverted double U variation with a first 65.2% after 4 years peak at 3 years (15.2%) and a second 45.6% after 6 years peak after 7 years (35%) 33.7% after 7 years USA and global USA and global USA and global ( ) ( ) (Summer Spring 91) 85.3% after 2 years 54% after 4 years 41.2% after 6 years 33.4% after 10 years Inverted double U variation with a first peak at 4 years (20.9%) and a second peak at 6 years (15.1%) Differentiation of Survival rate*** Lower survival of international joint ventures - No differences in survival rate 14.5% after 15 years - No differences in survival rate 85% after 2 years 69% after 4 years 53% after 6 years 30% after 10 years Global % after 2 years 38.8% after 4 years 31.4% after 6 years 23.1% after 10 years - - Inverted double U variation with a first peak at 1 year (20.7%) and a second peak at 6 years (25%) No differences in survival rate * The period between parentheses corresponds the reference period during which announcements of formation of joint ventures were gathered ** The annual termination rate is equal to the ratio of (a) the number of joint ventures that ended during the studied year and (b) the total number of joint ventures that survived *** This column presents results concerning the existence of differences in survival rate between domestic and international joint ventures

10 2. Research Methodology: Sample, Survival analysis and Variables This study is based on a sample of 139 joint ventures involving internationalization or expansion into emerging countries, including at least one partner from countries of the European Union and for which the announcement of their foundation was made between January 1996 and December These announcements were gathered and cross-referenced from databases (of Les Echos, The Financial Times, The Wall Street Journal and LexisNexis) and business reports of the European companies involved in these joint ventures. These different sources of information were also used for keeping track of the development of these joint ventures from their date of creation during year 1996 until June Throughout this period, we identified joint ventures that survived and those that had ended. Joint ventures were considered as ended when they were dissolved, sold off to one of the partners, completely sold off to a third party or acquired by one of the partners. A total of 32 internationalization joint ventures were dissolved, sold off or acquired before the end of the period analyzed. Table 2 shows, year by year during the entire period analyzed, the number of joint ventures that were dissolved, sold off or acquired, and thereby calculate the annual termination rate of joint ventures in this sample. Table 2 Annual variation of the termination rate of joint ventures Surviving joint ventures Ended joint ventures Annual termination rate 0% 5% 5,3% 4% 5% 3,5% 2,7% The analysis of annual termination rate of internationalization joint ventures reveals an identical trend to that observed in research literature (Kogut, 1988, 1989 et 1991; Park & Russo, 1996; Shi, 1998): an inverted double U variation pattern is observed with two specific years (the third and fifth) during which there is a greater than normal increase in this rate.

11 The joint venture duration, which is the dependent variable of this study, was measured in numbers of years. For non-censored joint ventures i.e., those that ended before the end of the period analyzed the duration is deduced by calculating the number of years between the creation and the end of the joint venture. For censored joint ventures i.e., those that did not end before the end of the period analyzed the duration is also equal to the number of years that passed between the creation and the end of the period analyzed. Therefore the maximum duration of joint ventures in the sample is 7.5 years. By statistical analysis of survival, it is possible to derive a model of the duration or longevity of joint ventures, by estimating their hazard rate. The hazard rate is defined as the probability that an individual would experience an event in an interval from time t to t+s, given that the individual is at risk from time t (Allison, 1984: 23). Figure 1 shows the annual variation in the hazard rate of joint ventures. Hazard Rate 0,075 Figure 1 Annual variation in the hazard rate of joint ventures 0,05 0, Duration of Joint Ventures (Year) Figure 1 shows non-uniform variation in the hazard rate of internationalization joint ventures. This figure shows that the probability of termination of internationalization joint ventures is low in the first two years of their existence and only becomes significant from the second year to the fourth year. This observation is in agreement with that of Park & Russo

12 (1996) and Hennart et al. (1998), who identified a honeymoon effect in their sample of joint ventures. This honeymoon effect was first established by Levinthal & Fichman (1988) after survival analysis of relations between audit companies and their clients. In the more general discussion of the survival of organizations, this effect is opposed to the liability of newness, which is put forward by the population ecology of organizations school (Caroll & Delacroix, 1982; Freeman et al., 1983), which itself refers to a very high hazard rate in the very first years of an organization s existence. In order to explain variations in the hazard rate of joint ventures, independent variables were introduced into a regression model. However, it was firstly necessary to find the regression model that had the best adjustment of hazard rate distribution. Four models were tested Weibull, exponential, log-logistic and lognormal and their Chi-square values were compared in order to retain only the best. The log-logistic model, characterized by an inverted U non-uniform variation, obtained the best adjustment of distribution of the hazard rate of internationalization joint ventures (Chi-square = 15.83, p < 0.05). The trend observed in Figure 1 for annual variation in the hazard rate is thus confirmed by the Chi-square selection test. The results of this test show that, when explanatory variables other than time (i.e., the joint venture duration) are introduced into the regression model, it is always a nonuniform (i.e., log-logistic) model, and not a uniform model (i.e., Weibull or exponential), that obtains the best adjustment. Two categories of independent variables were introduced into the regression model: first of all, explanatory variables measuring the country risk, then a set of control variables. There are three explanatory variables of country risk, but they all come from the same source, i.e., the Euromoney database. A first variable is the overall country risk score. This variable was measured at the end of the analyzed period (for censored joint ventures) and on the date of the end of the alliance (for non-censored joint ventures ). The overall country risk score proposed

13 by Euromoney corresponds to a weighted mean of nine indicators, of which the most important are political risk (25% of the total score), economic performance (25%) and the debt size (10%). This score is positioned on a scale ranging from 0 (maximum risk) to 1 (no risk). A second variable is variation in country risk, which corresponds to variation in overall country risk score between the period of creation of the joint venture and the end of the analysis period (for censored joint ventures) or the date of the end of the alliance (for noncensored joint ventures ). A last variable, deterioration of country risk, was developed by cross-checking the variation in country risk and the overall country risk score on creation of joint ventures in This variable only covers a part of the sample because it only concerns joint ventures that have experienced negative variation of country risk during the period analyzed (i.e., 116 joint ventures out of a total of 139). This is an ordinal variable composed of three categories of joint ventures. Category 1 corresponds to joint ventures formed in countries with high country risk (overall score less than 0.6); category 2 corresponds to joint ventures formed in countries with average country risk (overall score between 0.60 and 0.75); and category 3 to joint ventures formed in countries with low country risk (overall score of more than 0.75). Table 3 presents the distribution of the three country risk variables in the sample. Table 3 Statistical description of country risk variables Overall Country risk score Average = 0.56 Min. = 0.26 Max. = 0.9 Min. = Max. = 0.67 Mean deviation = 0.11 Variation in Country risk Average = Mean deviation = 0.18 Deterioration in Country risk Joint ventures % Category ,3 Category ,2 Category ,5 Total At the end of the analyzed period, the analysis of the average overall country risk score reveals a relatively high country risk for the joint ventures concerned. Table 3 also shows that

14 joint ventures of the sample were formed in countries where the average country risk deteriorated during the period analyzed (i.e., ). It must be remembered that this period was marked by major regional crises that resulted in a sometimes very marked deterioration in country risk score, particularly Asian, Russian and Latin American crises. An analysis limited to only those joint ventures that experienced negative variation in country risk during the analyzed period shows that more than two thirds of joint ventures belong to categories 2 and 3, i.e. they concern emerging countries whose country risk was low or very low in 1996 and then greatly deteriorated a few years later. The control variables are: inclusion of a local State partner (0-yes, 1-no), line of business (0-industry, 1-services), geographic area (1/ Asia, 0-yes, 1-no; 2/ Latin America, 0-yes, 1-no; Central & Eastern Europe, 0-yes, 1-no) of the joint venture; and equity ownership between partners (0-50/50, 1-unequal equity ownership). Table 4 shows the distribution of control variables in the sample. Table 4 Statistical Description of Control Variables Inclusion of a local State partner Joint ventures % Yes No Total Line of business Joint ventures % Industry Services Total Geographic area Joint ventures % Asia Yes No Latin America Yes No Central & Eastern Europe Yes No Total Equity ownership Joint ventures % 50/ Unequal equity ownership Total

15 By studying Table 4, it is possible to identify the following trends within the sample: (1) The local partners involved are mainly private companies; (2) Joint ventures are mainly formed for industrial purposes and in the Asian area; (3) Most joint ventures are of the 50/50 type. 3. Analysis of the Impact of Country Risk on the Survival of Internationalization Joint Ventures Table 5 presents a matrix of correlations between the different quantitative variables that are used in the log-logistic regression model. Leaving aside marked correlations involving one or all of the three dummy variables of geographic area (i.e., Asia, Latin America, and Central & Eastern Europe), Table 5 shows that the strongest correlation is that associating (a) the inclusion of a local public partner within joint ventures and (b) the survival of joint ventures (r = -0.25). In other words, the correlation matrix of does not show any particular problems of colinearity between explanatory variables.

16 Variables Table 5 Correlation Matrix Joint venture survival a 2. Overall country risk score.19** 3. Variation in country risk Local State partner -.25***.15.21** 5. Business activity Geogr. dummies Asia ***.61*** *** 7. Latin America *** *** *** 8. Eastern Europe.007.2** -.64***.26*** *** Equity ownership * p <.1 ** p <.05 *** p <.01 (two-tailed tests) a «Joint Venture Survival» is a dummy variable (0-yes; 1-no). This variable was used to obtain the duration and hazard rate of joint ventures.

17 Table 6 presents four log-logistic analyses of the impact of different country risk variables on the hazard rate of internationalization joint ventures. The first analysis (i.e., model 1) corresponds to a limited regression analysis, whose only explanatory variables are the overall country risk score and the variation in country risk. The second model jointly tests the impact of the two preceding country risk variables and that of all the control variables (inclusion of a local State partner, line of business, geographic area Asia, Latin America and Central & Eastern Europe and the sharing of equity ownership between partners). By including the country risk deterioration variable, the third model proposes an analysis of the impact of country risk limited to only joint ventures that experienced negative variation in this risk during the period analyzed. The last model (i.e. model 4) addresses this sub-sample of joint ventures and jointly tests the impact of the three country risk variables and that of all the control variables. Table 6 Log-logistic regressions a, b Variables Model 1 Model 2 Model 3 Model 4 Intercept 1.934*** 1.906*** 1.798*** 1.745*** Country risk global score *.278** Country risk changes Country risk build-up *** -.039* -.317*** -.06*** Local State partner.027**.033*** Business activity Geogr. dummies Asia Latin America Eastern Europe Equity ownership.023**.021* Model chi-square * *** *** Incremental chi-square * ** Wald ratio Likelihood ratio ** * *** *** *** *** N a Positive coefficients indicate that increase in independent variable values increase the duration of joint ventures and decrease their hazard rate. b Cell entries are coefficient estimates. * p <.1 ** p <.05 *** p <.01 (two-tailed tests) It is interesting to note that country risk variables show no significant impact in models 1 and 2, but they are determinant factors in models 3 and 4. These variables are most important in the case of internationalization joint ventures formed in countries that accumulate country

18 risk during the period analyzed. It is solely in this context that the joint venture survival is significantly affected by the country risk score and its variations. The relation between the joint venture survival and country risk is only linear in this particular context. In a very different context, such as that of countries where the risk is decreased, no significant relation was observed between these variables. An in-depth study of models 3 and 4 shows the nature of the impact of country risk variables on the joint venture survival. Therefore joint venture survival is threatened in countries that traditionally high risks, and which undergo marked deterioration of this risk as time progresses. Again in this particular case of countries that accumulate country risk, the converse relation is valid: i.e., the hazard rate of joint ventures tends to decrease in low-risk countries that only undergo very minor deterioration of this risk as time progresses. As distinct from country risk, the control variables have very stable results, irrespective of the model analyzed. Among these control variables, only two have a significant impact on survival: (i) inclusion of a local State partner within the joint venture, and (ii) ownership of the joint venture between the partners. More particularly, Table 6 shows that the inclusion of a local State company among partners and a 50/50 equity ownership of the joint venture s capital could increase the joint venture s hazard rate. The negative impact of the inclusion of a local State partner is a surprising result, particularly when one lists the specific resources that can be provided for the joint venture by a local State partner (ease of access to State contracts, to cheap capital, and to people in charge in State authorities, etc.); but it complies with the observations made by other studies (Raveed & Renforth, 1983). The negative impact of 50/50 equity ownership is a result that is part of a classic debate, started in the 1980s (Killing, 1983; Gomes-Casseres, 1987; Shenkar & Zeira, 1990; Bleeke & Ernst, 1991) and which aims to show that the sharing of equity ownership is one of the main factors that determines the joint venture stability or instability. This result is identical to that obtained by Blodgett (1992),

19 Park & Ungson (1997) and Meschi (2003), who reached the conclusion that sell-offs and dissolutions of joint ventures appear to be more important and more significant in 50/50 type joint ventures than in joint ventures with unequal equity ownership. Conclusion This paper had two aims: to examine the survival of internationalization joint ventures formed in emerging countries, and to relate their hazard rate to variations in the country risk of the emerging countries concerned. The decision to address these questions was motivated by the idea that joint ventures formed for internationalization or for expansion into emerging countries are apparently more unstable than other types of joint ventures, and that country risk is a major explanatory factor of this instability. This idea is based on the results of various research studies (Raveed & Renforth, 1983; Lee & Beamish, 1995 ; Kale & Anand, 2001) that were summarized by Yan (1998: 773) as follows: for international joint ventures formed in developing or transforming economies, the turbulent political and economic environments together with the intercultural and interorganizational dynamics have made managing international joint ventures particularly challenging. The results obtained from the survival analysis question the specificity of internationalization joint ventures as regards stability, and more precisely duration. Concerning the survival of internationalization joint ventures formed in emerging countries, time-related variations observed in their termination rate (i.e., inverted double U variation) and their hazard rate (i.e., curvilinear distribution with honeymoon effect) are identical to those revealed by research literature (Kogut, 1988, 1989 et 1991; Park & Russo, 1996; Hennart et al., 1998; Shi, 1998) for different types of joint ventures (domestic joint ventures, international joint ventures formed in developed countries, etc.). To sum up, the agreement between our results and those in existing research literature leads us to refute the idea of such

20 a specificity of the time-related variation of joint ventures that are formed for internationalization or for expansion into emerging countries. What is the role played by country risk in the instability and the outcome of internationalization joint ventures? In the context of emerging countries, country risk is part of these explanatory variables that are normally considered to be critical for the survival of joint ventures (Yan, 1998). However, its impact is not always empirically confirmed in every case. Questions remain concerning the real impact of country risk and its variations in time, but also concerning the nature (uniform/non-uniform, continuous/discontinuous) of this impact. The results obtained provide only a partial confirmation of the impact of country risk: the relation between country risk and the joint venture survival is confirmed empirically, but only when the emerging country concerned undergoes deterioration of its risk as time progresses. The relation between these two variables is discontinuous. It is confirmed and negative in the case of countries that accumulate risk, but disappears in the case of countries whose risk is minor. By trying to identify the role played by the country risk, this paper contributes to the knowledge of factors that explain the termination of joint ventures. Of course, the outcome of joint ventures could not be explained exclusively by the country risk and its variations. However, this study shows that, among the economic and political factors affecting the survival of joint ventures which were grouped together under the expression unexpected local contingencies by Yan (1998: 778), the country risk is an important factor. In the continuation of this research, it would be interesting to analyze the other major factor belonging to this category of unexpected local contingencies, that is, local legislation concerning direct foreign investment and the easing of its restrictions as time progresses, and to compare its impact on the survival of joint ventures with the impact of country risk. Like country risk and its variations, local legislation concerning direct foreign investment and the easing of its

21 restrictions are considered to be critical for the survival of joint ventures. However, here again, the empirical confirmation of this relation is questioned, between, on one hand, studies that confirm the existence of such an impact (Kale & Anand, 2001) and, on the other hand, studies that find no such impact (Prevot & Meschi, 2003). References Barkema, H. G. & Vermeulen, F. (1997). What differences in the cultural backgrounds of partners are detrimental for international joint ventures? Journal of International Business Studies, 28(4), pp Bleeke, J. & Ernst, D. (1991). The way to win in cross-border alliances. Harvard Business Review, 69(6), pp Blodgett, L. L. (1992). Factors in the instability of international joint ventures: An event history analysis. Strategic Management Journal, 13(6), pp Carroll, G. R. & Delacroix, J. (1982). Organizational mortality in the newspaper industries of Argentina and Ireland: An ecological approach. Administrative Science Quarterly, 27, pp Connolly, S. G. (1984). Joint ventures with third world multinationals: A new form of entry to international markets. Columbia Journal of World Business, Summer, pp Cox, D. R. & Oakes, D. (1983). The Analysis of Survival Data. Chapman & Hall, New York. Delios, A. & Beamish, P. W. (2001). Survival and profitability: The roles of experience and intangible assets in foreign subsidiary performance. Academy of Management Journal, 44(5), pp Elandt-Johnson, R. C. & Johnson, N. L. (1980). Survival Models and Data Analysis. John Wiley, New York. Franko, L. G. (1971). Joint Venture Survival in Multinational Corporations. Praeger, New York (NY). Freeman, J., Carroll, G. L. & Hannan, M. T. (1983). The liability of newness: Age dependence in organizational death rates. American Sociological Review, 48, pp Garrette, B. & Dussauge, P. (1995). Les Stratégies d Alliance. Les Editions d Organisation, Paris. Geringer, J. M. & Hebert, L. (1991). Measuring performance of international joint ventures. Journal of International Business Studies, 22(2), pp Gomes-Casseres, B. (1987). Joint venture instability: Is it a problem? Columbia Journal of World Business, Summer, pp Hamel, G. Doz, Y. L. & Prahalad, C. K. (1989). Collaborate with your competitors and win. Harvard Business Review, 67(1), pp Hamel, G. (1991). Competition for competence and inter-partner learning within international strategic alliances. Strategic Management Journal, 12(Special Issue), pp

22 Hennart, J.-F. (1988). A Transaction costs theory of equity joint ventures. Strategic Management Journal, 9(4), pp Hennart, J.-F., Kim, D.-J. & Zeng, M. (1998). The impact of joint venture status on the longevity of Japanese stakes in U.S. manufacturing affiliates. Organization Science, 9(3), pp Inkpen, A. C. & Beamish, P. W. (1997). Knowledge bargaining power and the instability of international joint ventures. Academy of Management Review, 22(1), pp Levinthal, D. & Fichman, M. (1988). Dynamics of interorganizational attachments: Auditorclient relationships. Administrative Science Quarterly, 33, pp Meschi, P.-X. (2003). Pourquoi et comment sortir d une alliance? Revue Française de Gestion, 143, pp Kale, P. & Anand, J. (2001). Effects of market liberalization on joint venture contributions, control, stability and performance: An empirical study of international joint ventures in India. Academy of Management Proceedings, Academy of Management Conference, Washington. Killing, P. (1983). Strategies for joint venture success. Praeger, New York (NY). Kogut, B. (1988). A study of the life cycle of joint ventures. Management International Review, 28(4), pp Kogut, B. (1989). The stability of joint ventures: Reciprocity and competitive rivalry. The Journal of Industrial Economics, XXXVIII(2), pp Kogut, B. (1991). Joint ventures and the option to expand and acquire. Management Science, 37(1), pp Lee, C. & Beamish, P. W. (1995). The characteristics and performance of Korean joint ventures in LDCs. Journal of International Business Studies, 26(3), pp Leung, W.-F. (1997). The duration of international joint ventures and foreign wholly-owned subsidiaries. Applied Economics, 29, pp Pan, Y. (1996). Influences on foreign equity ownership level in joint ventures in China. Journal of International Business Studies, 27(5), pp Park, S. H. & Russo, M. V. (1996). When competition eclipses cooperation: An event history analysis of joint venture failure. Management Science, 42(6), pp Park, S. H. & Ungson, G. R. (1997). The effect of national culture, organizational complementarity, and economic motivation on joint venture dissolution. Academy of Management Journal, 40(2), pp Pearce, R. (1997). Toward understanding joint venture performance and survival: A bargaining and influence approach to transaction cost theory. Academy of Management Review, 22(1), pp Prevot, F. & Meschi, P.-X. (2003). External evolution processes affecting joint venture: The case study of a French-Brazilian joint venture in the information technology sector. SMS Proceedings, SMS Mini-Conference, Buenos Aires. Raveed, S. & Renforth, W. (1983). State enterprise-multinational corporation joint ventures: How well do they meet both partner needs? Management International Review, 23(1), pp

23 Shi, Y. (1998). The effects of tasks on semiconductor start-up alliance dissolution: A structuration perspective. Journal of High Technology Management Research, 9(1), pp Yan, A. (1998). Structural stability and reconfiguration of international joint ventures. Journal of International Business Studies, 29(4), pp Zeira, Y. & Shenkar, O. (1990). Interactive and specific parent characteristics: Implications for management and human resources in international joint ventures. Management International Review, 30(Special Issue), pp