In this chapter, there are three sections. In the first section, the statistical results will be explained. In the second section, we will discuss the four hypotheses respectively according to the empirical results. Finally, the limitations to this study are in the third section.
4.1 Results
Concerning descriptive statistics, Table 1 presents the correlation coefficient matrix sampling information such as the average value of individual variables and standard deviation.
Because the correlation coefficient matrix indicates that correlations between some variables are greater than .50, suggesting a correlation that is higher than usual, we used VIF to test multicollinearity in the analysis process. Although a number of variables are correlated with other variables at the 0.05 level, the largest variance inflation factor (VIF) in the model (2.178) is far below 10, and the mean VIF value (1.426) is close to 1, suggesting that multicollinearity does not threaten the validity of our coefficient estimate.
Table 2 presents the empirical results. The dependent variable is the growth rate of employees in the second time period (2005-2007). The key explanatory variable (PREGROW) is the growth rate of the employee in the preceding time period (2001-2004). The hypothesis 1 predicts that the Penrose effect would present itself. As shown in Model 2, the coefficient of PREGROW is negative and statistically significant. This finding provides empirical support for hypothesis 1 that a fast growing firm might not maintain its high rate of growth and might experience a deceleration of growth in the subsequent time period.
Hypothesis 2 purposed that through alliances, firms could overcome the managerial limitations and maintain or improve on growth rates. As for the study’s findings on the allied firm's ability to overcome the Penrose effect, we did not find any empirical evidence supporting
32
the assertion that firm can overcome the Penrose effect through alliances. The coefficient of NUMALLY in Model 3 is positive but not statistically significant, suggesting that the number of alliance formed by firms cannot predict the growth rate of the allied firms in the subsequent periods. Therefore, hypothesis 2 is not supported. We speculate that the benefits of having more alliance partners cannot offset the costs associated with alliance management.
The above results confirm the importance of capabilities to the allied firms suggested by hypothesis 3 and hypothesis 4. Alliances are not always successful. Prior research
demonstrated that firms with capabilities are more likely to succeed and create value from an alliance than are firms lacking these capabilities. Hypothesis 3 and hypothesis 4 asserted the alliance capability and contracting capability separately.
Hypothesis 3 discussed about alliance capability and asserted that the more previous alliance experiences a firm has, the more a firm can overcome the managerial limitations to firm growth through alliance. We used the total number of previous alliances in the previous time period as measurement. The coefficient of PREALLY in Model 3 is positive but not statistically significant, suggesting that the number of alliance formed by firms in the previous time period cannot predict the growth rate of the allied firms in the subsequent periods.
Therefore, hypothesis 3 is not supported. We speculate that in dealing with inter-firm
relationship, having only trust cannot help firms to overcome the managerial limitations and maintain or improve on growth rate.
Hypothesis 4 discussed about contracting capability and purposed that firms with stronger contracting capability could grow faster. Contracting capability comes from previous
contracting experiences. We use the total number of litigations in the previous time period as measurement. As shown in Model 3, the empirical results indicate that the coefficient of NUMLIT is positive and statistically significant at the, providing empirical support for the
33
assertion of firms with contracting capability can help firms to overcome the managerial limitation and grow faster in the subsequent time period. Therefore, the hypothesis 4 is empirically supported.
4.2 Discussions
There are many studies examine how much a firm grows. However, the growth literature rarely discussed how firms grow. To focus on this issue, there are three main purposes of this study. The first one is to examine contractual organizational forms, especially strategic alliance, and its implications for firm growth. Instead of measuring the growth rate, we focus our interests on “how” aspect of firm growth. Also, this study divided capabilities into alliance capability and contracting capability, and then examined the importance of each capabilities and their impact on firm growth. The last purpose in this research is to emphasize on the contracting capability and its implications to firm growth.
According to Penrose (1959), firms either organically or acquisitionally grow. Yet, there is adjustment cost when firms grow though merger and acquisition. Therefore, Penrose suggested firms should grow organically. However, firms face managerial limitations to organic growth.
In order to overcome the managerial limitations and grow faster, firms may use contractual organizational forms to grow (Norton, 1998; Shane, 1996; Teece, 1986). This study goes on with growth literature, and responses to the “how” issue instead of “how much” issue. We examine whether contractual organizational form, especially strategic alliance, can help firms overcome the managerial limitations and can facilitate the growth rates or not. However, prior researches haven’t examined the relationship between alliance and firm growth. We suggest this type of (contractual) organizational form will provide the resources needed for firm growth while undertaking less adjustment costs of recruiting new employee and/or agency costs of
34
monitoring staffs. Alliances provide firms a way to find resources and to recognize
opportunities. Through alliances, firms may gain complementary resources without incurring costs related to further managerial investment. Alliances also help firms recognize
opportunities through interflows of novel information. This form of assistance saves managers both time and energy searching for and identifying complementary resources. Thus, the allied firms can leverage their existing managers' capacity to plan a subsequent growth program.
However, our empirical result is positive but not statistically significant, suggesting that the number of alliance formed by firms cannot predict the growth rate of the allied firms in the subsequent periods. We speculate that there may be several reasons lead to this result. First, from the perspective of transaction cost theory, strategic alliance is sometimes risky due to contractual hazards (Williamson, 1975; 1985). Previous research also shows that firms
generally fail with half the alliances they had formed because alliances are difficult to manage (Blaeeke & Ernst, 1993; Kogut, 1988; AllianceAnalyst, 1996). According to Wittmann (2007), alliances are more likely to fail when: (1) firms choose to form alliances based on modeling competitor’s actions; (2) firms do not systematically prioritize alliances and allocate resources in line with strategic needs. However, there are few discussions about adjustment cost in strategic alliance literatures. Therefore, this study assumes that firms can grow through the use of strategic alliance because there is no adjustment cost. Yet, this study finds out that there is still some costs while firms using contractual organizational forms to grow. We suspect that strategic alliance will lead to the change of process and routine within organization due to interfirm learning and imitation. Therefore, strategic alliance might produce adjustment cost if the allied firms invest in partner-specific resources and/or adopt partners' practices. However, some change will enhance the capacity for innovation and growth, while some will produce more conflicts. Therefore, we purpose that firms should choose the form of strategic alliance carefully and cultivate the capability to deal with the adjustment cost while using strategic
35
alliance as growth strategy.
Secondly, prior researches have demonstrated the difficulty in managing alliances and the failure of alliances. This study further suggests that there is adjustment cost while using strategic alliances as growth method. Therefore, the benefits of having more alliance partners cannot offset the costs associated with alliance management. We then use regression method to examine the effect of squared number of alliance. The result shows that squared number of alliance and growth are negatively (-2.726; p=.010) related. Therefore, we argue that there may be an optimal scale of alliances and that the relationship between the number of alliances and firm growth may form an inverted-U shape. This result indicates that since alliances are difficult to manage, the more alliances a firm concludes, the more costs and complex problems the firm may face. Therefore, it is important for firms and managers to manage the alliances portfolio properly.
Thirdly, prior literatures on strategic alliance have demonstrated that strategic alliance brings resources and growth opportunity into organization. Therefore, this study combines strategic alliance and firm growth, and suggests that strategic alliance can help firms grow.
However, the empirical result in this study does not statistically support our theory. In addition to the prior speculations, another important reason is that what strategic alliances bring into organization is not the number of employee, but the managerial capability. If firms have more opportunities to form strategic alliances, their experiences might be internalized as managerial routine and capability. Therefore, firms with alliance experiences tend to leverage alliance capability to access to external resources thus streamline the size of the firm, in particular the size of employee.
Firms need to develop capabilities in managing alliances. We identify two kinds of capability that facilitate firm growth, including alliance capability and contracting capability.
36
However, the nature of capability differs in its implications on the growth of the firm.
Alliance capability is a kind of behavioral capability which is based on trust among partners.
On the contrary, contracting capability is firm-specific capability that is based on firms' knowledge about legal contracts. Both alliance capability and contracting capability can be developed by previous alliance experiences. The empirical results in this research shows that the impact of alliance capability on firm growth is negative, while the impact of contracting capability on firm growth is significantly positive. Therefore, we speculate that there may be some substitute relation between those two capabilities. Bardach and Eccles (1989) and Powell (1990) have stated the substitute relationship between trust and contractual guards. Our study confirms this relationship and suggests that development of formal governance has positive impacts on firm growth.
There may be some reasons that trust in alliances cannot facilitate firm growth. First, interfirm trust is a kind of specific, especially partner-specific asset. Trust that is attached on one particular partner is less likely to generate a spillover effect that promotes firm growth.
Trust can work under the circumstance that two companies have had aligned with each other before. However, this kind of relationship cannot be transfer from one partner to another partner.
Therefore, when firms need to enter new alliance and find new partner, trust accumulated in previous alliances doesn't work. As a result, firms cannot facilitate its growth rates by having this alliance capability based on interfirm trust.
On the other hand, contracting capability is firm-specific but less partner-specific capability. It can be redeployed and be easily transferred from partners to partners. Having contracting capability, firms may apply this general capability into new alliances. Thus, the more contracting capability a firm has, the more it can help firms to grow faster. The empirical result in this study also supports this perspective.
37
As to the contracting capability, this study uses the total number of litigation as variable to measure contracting capability. And the empirical result shows that firms with more
litigation experiences are more likely to maintain high growth rates in successive time period.
However, this result doesn’t suggest that firms should produce as much conflicts as they can in order to maintain high growth rates. There are more or less conflicts and disputes in the alliance partnerships between firms. One of the functions of contracts is to lessen and to resolve the disputes. However, contract incompleteness only arose after conflicts and disputes.
Sometimes conflicts and disputes can be solved internally. While sometimes, arbitrations or litigations are necessary. Only after actually experiencing litigation did the firms learn the contract incompleteness and attempt to make adjustment in the later contracts. In order to solve such disputes, or to prevent similar disputes in the future, firms make adjustment or add clauses in the later contract. Argyres and Mayer (2004) found that the uses of arbitration clauses were increasing over time. This study also suggests that in the process of litigation, firms may find the lacks of contract and learn how to sign a more complete contract in the future. Therefore, firms that experienced more litigation would have accumulated more contracting capability and are more likely to maintain high growth rates in successive time period.
4.3 Limitations
There may be some limitations to this study. First, its sample size might have distorted the statistical results. If this study amplifies the sample size, the effects that the number of alliances can have on firm growth may be empirically significant. Second, the measurements of alliance and alliance capability are the total number of alliances in 2004 and the total number of alliances in the previous time period (2001-2003). All of the data collected in this study are
38
from the news released by the firms in TEJ database. The news only posted the significant events. There may be some alliances between firms were unreleased in the news. Therefore, the total number of alliances may be underestimated. Third, the measurement of contracting capability is the total number of litigations in the previous time period (2001-2003). Since contracts are secrets to firms, it cannot be revealed or easily read publicly. Therefore, this research uses the number of litigation as the measurement to measure contracting capability.
Only when there is litigation, the problems of contracts can be revealed. The same limitation with the third, all of the data collected in this study are from the news released of the firms in TEJ database. The news only posted the significant issue. Therefore, the real total number of litigations may also be underestimated.
As to Penrose effect, although our empirical results indicate that firms’ possession of alliance capabilities may help the firms improve their growth rate in the next time period, the results do not constitute statistically compelling evidence that using alliances and possessing alliance capabilities can reduce the Penrose effect. We suggest that future research examine, on the basis of this empirical study, whether or not an optimal scale and a curve relationship exist.
Also, because (1) this study may have other limitations such as routine inconsistency and resource heterogeneity, (2) the Penrose effect is complicated, and (3) the mechanism for lessening the Penrose effect varies across firms and industries, we suggest that future research might broaden the governance and capability-based view by considering multiple forms of contractual organization that concern the relationships between markets and hierarchies.
39