THEORY AND HYPOTHESES
- Top of page
- THEORY AND HYPOTHESES
- ADDITIONAL ANALYSIS: POST-ACQUISITION PATENTING
- ROBUSTNESS TESTS
- DISCUSSION AND CONCLUSION
In innovative industries, technological change is rapid and frequent (Sarkar et al., 2006). While both incumbents and start-ups strive to innovate, past research provides evidence that much of the truly novel innovations originate in start-ups (Abernathy and Utterback, 1978; Pavitt, Robson, and Townsend, 1987). As a result, technological acquisitions have become a popular complement to internal innovation, allowing firms to overcome the time compression diseconomies (Dierickx and Cool, 1989) inherent in the rapid and frequently changing technologies of innovative industries.
Accordingly, technological acquisitions have become an important stream in the broader acquisitions literature (Ahuja and Katila, 2001; Benson and Ziedonis, 2009; Capron and Mitchell, 2009; Graebner, 2004, 2009; Kapoor and Lim, 2007; Makri et al., 2010; Paruchuri et al., 2006; Puranam, Singh, and Chaudhuri, 2009; Puranam, Singh, and Zollo, 2006; Puranam and Srikanth, 2007; Ranft and Lord, 2002; Schweizer, 2005). Work in this stream has focused on the acquisition of small, technology-intensive target firms, as do we. Beyond their managerial importance, such acquisitions allow researchers to focus on the effects of technological synergies by minimizing the impact of potential confounding factors usually present in the acquisition of large and/or nontechnological targets, such as cost (scale) synergies or market power synergies.
Barney (1988) argued that acquirers can capture economic value by creating novel recombinations from their resources and capabilities and those of the target. More recent research provides evidence that the pursuit of such novel recombinations motivates many acquisitions (Karim and Mitchell, 2000; Larsson and Finkelstein, 1999). A consistent finding is that the degree to which the technological knowledge bases of the acquirer and target overlap (i.e., technological overlap) affects the acquirer's ability to generate novel recombinations.
However, the measurement of technological overlap in existing research obscures important contingencies that determine the potential to create or even destroy value post-acquisition. Prior studies have measured technological overlap as the amount of the target's knowledge base that the acquirer already knows (Ahuja and Katila, 2001), the sum of the technological overlap from both firms (Mowery, Oxley, and Silverman, 1996, 1998), or subsumed technological overlap in a more coarse-grained measure of resource overlap represented by product lines and product categories (Karim, 2006; Karim and Mitchell, 2000). Examining target and acquirer overlap separately allows us to explore the possibility that they differentially influence the ability of the firm to create value from its own technological capabilities and those of the target. We focus on technological capabilities as the source of potential value because as Grant (1996: 380) stated, ‘the key to sustainable advantage is not proprietary knowledge itself, but the technological capabilities which permit the generation of new knowledge.’
Our research question demands that our hypotheses examine the interaction between overlap and technological capabilities, allowing us to determine how, for example acquirer overlap affects the value created by acquirer technological capabilities, separate from its effect on the value created by target technological capabilities. The resulting hypotheses tell us whether the contribution of a ‘unit’ of technological capability to value creation increases or decreases as technological overlap increases. That is, they tell us how effectively the firm translates technological capabilities into value. Incorporating both overlaps allows us to test simultaneously the effects of absorptive capacity, knowledge redundancy, and post-acquisition conflict in a way prior studies have not been able to accomplish.
Our measure of value creation is cumulative abnormal return of the acquirer's stock. This measure has been extensively used to evaluate the value created from an acquisition in the management literature (Arikan and Capron, 2010; Capron and Pistre, 2002; Goranova, Dharwadkar, and Brandes, 2010; Kim and Finkelstein, 2009; McWilliams and Siegel, 1997; Uhlenbruck, Hitt, and Semadeni, 2006). Although used less often in the study of technological acquisitions, it offers a special advantage in this context. It captures investors' perceptions of the acquirer's ability to create future cash flows from unique and potentially inimitable synergies generated by recombining its technological capabilities with those of the target, which Barney (1988) has identified as the source of acquirer value appropriation in acquisitions. Since our interest is in the degree to which overlap shapes the realization of such synergies, cumulative abnormal return (CAR) offers benefits beyond other common measures, such as patent counts, in capturing not only the amount of subsequent inventive activity but also whether that activity represents novel synergies likely to generate value.
Figure 1 illustrates four idealized combinations of target overlap and acquirer overlap. The squares represent each firm's knowledge base, that is, the set of knowledge a firm has ‘demonstrated familiarity with, or mastery of,’ as described by Ahuja and Katila (2001) in their seminal paper. Citing Kim and Kogut (1996), Ahuja and Katila further describe the knowledge base as ‘the distinct elements of knowledge with which the firm has revealed a relationship.’ The shaded area represents technological overlap, that is, knowledge common to both target and acquirer. The nonshaded areas represent knowledge unique to the acquirer or target.
Figure 1. Asymmetries in acquirer and target technological overlap. A = Acquirer's technological knowledge base; T = Target's technological knowledge base
Download figure to PowerPoint
We begin by considering the impact of target overlap. When target overlap is low (quadrants 1 and 2), a large proportion of the target's knowledge is new to the acquirer and there are many opportunities for novel combinations of the target's knowledge and the acquirer's knowledge. However, the acquirer may not be able to realize these novel recombinations because it lacks absorptive capacity (Cohen and Levinthal, 1990). Lacking what Zahra and George (2002) call potential absorptive capacity, the ability to value and acquire external knowledge, and realized absorptive capacity, the ability to transform and exploit external knowledge, the acquirer will be unable to extract maximal value from the target's capabilities (cf. Mowery et al., 1998). Poor performance may occur because the acquirer is unable to incorporate and exploit the target's capabilities and/or because the target never possessed the amount or type of capabilities the acquirer believed it did.
When target overlap is high (quadrants 3 and 4), the acquirer has a greater ability to understand and absorb the target's knowledge. However, since much of the target's knowledge is redundant to knowledge the acquirer already possesses, there are fewer possibilities to create novel recombinations of the target and acquirer's knowledge. Indeed, the increase in knowledge redundancy not only decreases the possible number of novel recombinations that can be made using the newly acquired knowledge; it may also decrease the novelty and quality of those recombinations (Makri et al., 2010).
In summary, when target overlap is low, there are many opportunities for novel recombinations, but the acquirer lacks the absorptive capacity to recognize and execute them. The technological resources of the target are largely wasted. When target overlap is high, the acquirer has the necessary absorptive capacity, but knowledge redundancy means there are few novel recombinations available. The target's capabilities offer few opportunities to create value.
If both the positive impact of absorptive capacity and the negative impact of redundancy increase linearly as target overlap increases, their combined effect will, of course, be constant—a given change in target overlap will change the value created by the target's capabilities by the same amount at any level of target overlap. Whether the change is positive or negative depends on whether the impact of absorptive capacity or redundancy changes more quickly with changes in target overlap.
It seems more likely that absorptive capacity increases at a decreasing rate as target overlap increases. The negative impact of redundancy may also increase nonlinearly as target overlap increases, although it is less clear whether it will increase at a decreasing or increasing rate. The combination of these nonlinear effects is likely to be nonlinear as well. However, nonlinearity does not imply that the combined effect will necessarily be nonmonotonic when target overlap is between 0 and 100 percent, as it must be by definition.
Depending on the relative magnitude and curvature of the relationship between (1) value creation and absorptive capacity and (2) value creation and redundancy, their combination could yield three potential outcomes. Theory is uninformative as to the relative magnitude and curvature of these two relationships, so we offer competing hypotheses.
It may be the case that, for any increase in target overlap, the improvement in absorptive capacity increases the ability of the firm to create value from the target's capabilities by more than the increase in redundancy decreases it. Colloquially, familiarity outweighs novelty. The combined effect would be a monotonic, possibly nonlinear, increase in the ability of the acquirer to generate value from the target's capabilities as target overlap increases.
- Hypothesis 1a: An increase in target technological overlap will positively affect the impact the target's technological capabilities will have on abnormal returns.
It may rather be the case that, for any given increase in target overlap, the increase in redundancy reduces the ability of the firm to create value from the target's capabilities by more than the improvement in absorptive capacity increases it. Colloquially, novelty outweighs familiarity. The combined effect would be a monotonic, possibly nonlinear, decrease in the ability of the acquirer to generate value from the target's capabilities as target overlap increases.
- Hypothesis 1b: An increase in target technological overlap will negatively affect the impact the target's technological capabilities will have on abnormal returns.
The last possibility is that the combined effect is nonmonotonic as target overlap increases, with either a U or an inverted U-shaped relationship. The latter would result if, for a given increase over a low level of target overlap, the improvement in absorptive capacity increases the ability of the firm to create value from the target's capabilities by more than the increase in redundancy decreases it—a relationship that is reversed when starting from a high base.
- Hypothesis 1c: An increase in target technological overlap will have a nonmonotonic effect on the impact the target's technological capabilities will have on abnormal returns.
We next consider acquirer overlap, which this study is the first to examine. We hypothesize that acquirer overlap affects the value created by both the acquirer's and the target's capabilities. Increased acquirer overlap is associated with increased routine disruption and conflict between the knowledge workers of the target and acquirer, a recognized source of value destruction in the acquisitions literature (Mirvis, 1985; Paruchuri et al., 2006; Puranam et al., 2006).
When acquirer overlap is low (quadrants 1 and 3), there is little basis for conflict arising post-acquisition. Since the overlapping knowledge represents a small portion of the acquirer's knowledge base, few of the acquiring firm's knowledge workers will find themselves in competition with the target's knowledge workers. Rather, supplementing the acquirer's capabilities with complementary capabilities of the target can generate novel recombinations. Since both workforces can benefit from being associated with these new recombinations, there is incentive for cooperation, setting the stage to increase the value of both the acquirer's and the target's capabilities.
For example, in 2000 MKS Instruments acquired Applied Science and Technology (ASTeX), a supplier to the semiconductor industry that had low acquirer overlap. The Chairman and CEO of MKS Instruments stated that ‘ASTeX is an ideal strategic fit for MKS. We serve virtually the same markets with zero product duplication; our product lines are fully complementary… The combined technological capabilities of the two companies will enable us to add further value through new innovative product solutions’ (PR Newswire, 2000b). Because there was no significant corresponding activity preexisting within MKS, ASTeX was integrated as a distinct product group while simultaneously maintaining an effort to identify possibilities for companywide integration for new product development (Business Wire, 2001). The avoidance of reconfiguration and conflict allowed ASTeX to maintain its innovative capability post-acquisition, and it received two Semiconductor International 2002 Editor's Choice Best Product Awards for products developed by members of the ASTeX team at MKS (PR Newswire, 2002).
In contrast, as acquirer overlap increases, represented by a move from the left two quadrants to the right two quadrants in Figure 1, more of the acquirer's knowledge workers are redundant to, rather than complementary to, the target's knowledge workers. Since the target's knowledge duplicates a larger proportion of the acquirer's existing resources, the acquirer is unlikely to maintain two separate bodies of related expertise. Efforts to combine the acquirer's and target's knowledge bases will expose teams from both firms, as well as their respective routines, to disruption. The teams may not work well together due to differences in culture, norms, and routines for communication and problem solving, reducing their performance (Chatterjee et al., 1992). Further, the target's knowledge workers will go from being ‘big fish in a small pond’ to small fish in the large pond of the acquirer's preexisting capabilities, leading to lost standing and diminished productivity (Kapoor and Lim, 2007; Paruchuri et al., 2006).
Even if teams from the acquirer and target are not actually combined, their similar expertise makes it likely that they will find themselves competing for limited resources related to an already established capability rather than supplementing each other in order to build a new capability. As the firm seeks to resolve this internal competition, it is often the target's workforce that is reconfigured (Capron, Dussauge, and Mitchell, 1998; Capron, Mitchell, and Swaminathan, 2001; Capron and Pistre, 2002; Karim, 2006). The resulting loss of social status and centrality on the part of the target's innovators is one of the most significant drivers of lost technological productivity post-acquisition (Paruchuri et al., 2006). It often leads to increased turnover among the target's workers (Krishnan, Miller, and Judge, 1997), which is highly damaging to knowledge transfer after the acquisition (Ranft and Lord, 2000).
Sometimes, however, it is the acquirer's employees that bear the brunt of post-acquisition reorganization. This is particularly likely to be the case when the acquirer intends to upgrade or substitute their present capabilities with the superior capabilities of the target (Karim, 2006), as would be suggested by a firm acquiring a target that duplicates many of its existing capabilities, rather than relying on internal development to expand its capabilities (Capron and Mitchell, 2009; Helfat, 1994; Tripsas and Gavetti, 2000).
Whichever group of employees is ultimately most affected by the acquisition, employees are likely to resist the acquisition for fear of its potential negative effects on their careers (Greenwood, Hinings, and Brown, 1994; Walsh, 1988, 1989). Such resistance is well documented as a source of acquisitions failing to meet their expectations (Blake and Mouton, 1985; Hambrick and Cannella, 1993; Larsson and Finkelstein, 1999). Employee resistance impedes the ease of communication and post-acquisition interaction between the knowledge workers, which is critical for the successful creation of novel recombinations using the acquirer's and target's capabilities.
For example, in 2000 Baxter acquired North American Vaccines (NAV) to enhance its ability to be a leader in the vaccines market where it was at the time a strong regional player. There was significant acquirer overlap, meaning consolidation and reconfiguration were necessary to avoid duplication of activities already existing within Baxter—activities with which the NAV's corresponding activities could be consolidated. The accompanying rationalization of research efforts led to job cuts within former NAV research facilities within two years. Within three years, Baxter terminated independent research within NAV, consolidating efforts within their existing research operation to leverage better the combined proprietary technologies. While initial job cuts were concentrated within NAV, the consolidation exposed both Baxter and NAV's employees to significant disruption (PR Newswire, 1999, 2000a; Terry, 2003).
In summary, an increase in acquirer overlap means that more of the acquirer's entrenched knowledge workers face the risk of their knowledge being made redundant by the acquisition and thus have an incentive to resist the integration of the target's knowledge workers. The resulting conflict increases the likelihood of turnover, and the resulting loss of knowledge for both the target and acquirer and makes successful recombination of the acquirer's and target's capabilities more difficult. Thus, we expect greater acquirer overlap to increase conflict leading to a decrease in the acquirer's ability to extract value from the target's and the acquirer's capabilities. This leads to our second and third hypotheses.
- Hypothesis 2: An increase in acquirer technological overlap will negatively affect the impact the target's technological capabilities will have on abnormal returns.
- Hypothesis 3: An increase in acquirer technological overlap will negatively affect the impact the acquirer's technological capabilities will have on abnormal returns.
We have focused on acquirer overlap as a determinant of conflict. However, target overlap may also affect the degree of disruption experienced. This possibility doesn't affect our understanding of the relationship between acquirer overlap and disruption that drives Hypotheses 2 and 3 but may shape our understanding of the results from testing Hypothesis 1, for which we followed the existing literature's focus on absorptive capacity and redundancy. Thus, we briefly explore the potential relationship between target overlap and the degree of disruption experienced.
If disruption and conflict increased with target overlap, it would—like redundancy—have a negative impact as target overlap increased. Disruption and redundancy are therefore complementary explanations, and we cannot rule out that both help drive Hypothesis 1. As we discuss below, the overall pattern of our results is strongly consistent with the extant interpretation that redundancy is a major driving factor behind the negative impact of target overlap on value creation from the target's capabilities.
Further, we believe that the relationship between target overlap and the degree of disruption experienced is significantly weaker than that of acquirer overlap, at least in the setting of technological acquisitions. Larsson et al. (2001: 614, emphasis added) described acquisitions as, ‘the whole acquired firm being integrated with only the related parts of the acquiring firm. Being acquired will likely create disruption for all but those with Transitory career profiles.’ That is, all target employees, whether or not in the area of overlapping knowledge, will be exposed to disruptions due to changes in culture (Larsson and Finkelstein, 1999), imposition of the acquirer's managerial systems (Capron and Mitchell, 1998), disruption of organizational and social identity (Haunschild, Moreland, and Murrell, 1994), and reduction in high-powered incentives (Kapoor and Lim, 2007).
In contrast, Larsson et al. (2001) describe integration affecting ‘only the related parts of the acquiring firm.’ Acquirer employees related to the target's knowledge base are much more likely than other acquirer employees to experience disruption and conflict through changed status; disruption of identity as workgroups are shifted, merged, or otherwise reformed; an increase in nonroutine work related to integrating with target-firm employees; and the potential exit of colleagues who are displaced (or fear displacement) by target-firm employees. Life for unrelated acquirer employees will probably continue relatively undisturbed.
Thus, as acquirer overlap varies from low to high, the degree of conflict and disruption will also vary broadly from low to high. However, in light of the substantial, companywide disruption experienced by the target, the additional variation in disruption and conflict caused by the degree of target knowledge overlap is much more constrained.
ADDITIONAL ANALYSIS: POST-ACQUISITION PATENTING
- Top of page
- THEORY AND HYPOTHESES
- ADDITIONAL ANALYSIS: POST-ACQUISITION PATENTING
- ROBUSTNESS TESTS
- DISCUSSION AND CONCLUSION
Many prior studies of technological innovations have used post-acquisition patenting as a measure of performance (Ahuja and Katila, 2001; Cloodt et al., 2006; Kapoor and Lim, 2007; Makri et al., 2010). Therefore, we replicated our original Analysis III using the patent counts of the post-acquisition firm for the three years following the acquisition.
In four out of the six tests of our hypotheses, the two measures provide consistent results. Three results (Hypothesis 1 for low acquirer overlap, Hypotheses 2 and 3 for low target overlap) are identical in direction and significance. The test of Hypothesis 3 for high target overlap points in the same direction but loses statistical significance for patents (p = 0.140).
The first substantive difference regards Hypothesis 1 (how target overlap affects the contribution of target capabilities) when acquirer overlap is high. We found that an increase in target overlap strengthened the relationship between target capabilities and patenting but weakened the relationship between target capabilities and CAR. This contrast is consistent with the logic we developed for Hypotheses 1a and 1b.
High target overlap indicates that the firm is buying a capability pool it understands well. Per the previous literature and our logic for Hypothesis 1a, this should give the firm the absorptive capacity it needs to identify, value, assimilate, and commercialize successfully the recombinative possibilities of the target's resources (Cohen and Levinthal, 1990). Consistent with this logic, we find that the post-acquisition count of patents as a function of target capabilities increases with greater target overlap.
However, consistent with other previous literature (Ahuja and Katila, 2001; Makri et al., 2010) and our logic for Hypotheses 1b, if high target overlap occurs when acquirer overlap is also high, there are few possibilities to create truly novel recombinations of the target's and acquirer's knowledge. Thus, the innovations occurring in these acquisitions create only marginal added value. Accordingly, the target's capabilities provide less increase in CAR than they would if target overlap were lower and there were more opportunities for novel recombination.
Very similar logic explains the contrasting results for Hypothesis 2 (how acquirer overlap affects the contribution of target capabilities) when target overlap is high. We found that an increase in acquirer overlap increased the relationship between target capabilities and patenting but weakened the relationship between target capabilities and CAR. An increase in acquirer overlap may help target employees identify and execute more opportunities to combine its capabilities with the acquirer's capabilities post-acquisition. This increases the number of patents produced as a function of the target's capabilities. However, an increase in acquirer overlap compounds the redundancy created by high target overlap, meaning there are fewer possibilities to create truly novel recombinations. Thus, the target's capabilities provide less increase in CAR than they would given low acquirer overlap.
Combining the findings for patents and CAR suggests that the quantity and quality of post-acquisition innovation may be differentially affected by overlap, which echoes the findings of Makri et al. (2010) who found different predictors of invention quantity, quality, and novelty. It also illuminates the tension among Hypotheses 1a, 1b, and 1c. Increased target overlap increases the acquirer's absorptive capacity but diminishes the novelty of resulting innovations. Thus, the target's capabilities result in more patents but generate less value. Collectively, these results confirm both the appropriateness and the unique value of the CAR measure in studying technological acquisitions.
DISCUSSION AND CONCLUSION
- Top of page
- THEORY AND HYPOTHESES
- ADDITIONAL ANALYSIS: POST-ACQUISITION PATENTING
- ROBUSTNESS TESTS
- DISCUSSION AND CONCLUSION
This study advances our understanding of the performance of technological acquisitions by extending the concept of technological overlap and by incorporating the interaction of the level of technological overlap and the amount of technological capabilities possessed by the target and acquirer. In doing so, it makes four primary contributions.
First, it theoretically and empirically advances the concept of technological overlap by demonstrating that target overlap and acquirer overlap are separate constructs and need not be symmetric. Doing so complements other work (Makri et al., 2010) that takes a more multidimensional view of technological overlap. We believe that the idea of asymmetric knowledge overlap across partners is also applicable to alliances, extending insights generated by examining overlap as a symmetric attribute of an alliance (e.g., Mowery et al., 1996, 1998).
Distinguishing target and acquirer overlap enables the paper's second contribution, simultaneously considering the impact of overlap (both target and acquirer) and technological capabilities (again, both target and acquirer) on acquisition performance. This study is the first paper of which we are aware to do so. One of the paper's key insights, that target overlap and acquirer overlap have distinct effects on the acquirer's ability to create value from combining the target's and the acquirer's capabilities, comes directly from this advance.
The study's third contribution is to broaden the theoretical explanation of value creation in technological acquisitions by simultaneously incorporating three drivers: the acquirer's absorptive capacity, knowledge redundancy, and exposure to organizational disruption due to conflict between the acquirer's and target's knowledge workers. Drawing on prior papers that have discussed each driver in isolation, our separate measures of target overlap and acquirer overlap allow us to consider them simultaneously, even though we do not directly observe them.
The paper's fourth contribution is to extend the literature on technological acquisitions by studying both shareholder value creation and post-acquisition invention productivity (patenting) as outcomes. Although relatively neglected as a dependent variable for technological acquisitions, shareholder value creation provides theoretical and managerial insights. It is a direct measure of the degree to which the managers of the acquiring firm accomplish their ultimate goal of generating shareholder value by acquiring small technological firms. Our combined findings complement other studies of post-acquisition invention productivity (Ahuja and Katila, 2001; Cloodt et al., 2006; Kapoor and Lim, 2007; Makri et al., 2010; Paruchuri et al., 2006; Puranam and Srikanth, 2007). In particular, they suggest that the low innovation quantity observed in acquisitions with low target overlap may conceal an offsetting increase in the novelty and quality of innovations generated in such acquisitions. In this regard, it helps confirm Makri et al.'s (2010) finding that less quantity can be accompanied with greater quality when lower similarity and greater complementarity exists between the acquirer and target.
Three useful insights flow from these contributions. The first insight comes from our finding that, when target overlap is high, knowledge redundancy decreases an acquirer's ability to derive value from a target's capabilities, but when target overlap is low, there does not seem to be a negative impact from a lack of absorptive capacity. This finding suggests that acquiring managers may be better able to recognize what they do not know than recognize what they know too well. Complementing Coff's (2002) finding that potential acquirers are more likely to withdraw an acquisition when there is little knowledge overlap between the target and the acquirer, we expect management to follow through only on acquisitions with foreseen synergies. However, in pursuit of those potential synergies, they may get tunnel vision and not recognize excessive knowledge redundancy and its potential value destroying effects. As noted above, disruption and conflict may also play a role here.
The second insight regards the precedents and consequences of conflict between knowledge workers on value created by the target's capabilities. High acquirer overlap negatively affects the acquirer's ability to extract value from the target's capabilities only when there is simultaneously high target overlap. Even though much of the acquirer's knowledge is redundant, if it can be coupled with nonredundant knowledge from the target, the target's knowledge workers have much to offer the acquirer. This situation allows for synergy realization, which can provide a productive working environment instead of one of conflict and competition. However, when the target brings less new knowledge (high target overlap), there is less chance for complementarities and synergy realization. Without potential for complementarities between the two firms' knowledge workers, redundancies are likely to translate into a more competitive, hostile environment.
The third insight examines the precedents and consequences of conflict on value created from the acquirer's capabilities. Conflict resulting from high acquirer overlap can actually destroy the value of the acquirer's existing capabilities. Indeed, applying the estimates of Table 5 to an acquisition with mean levels of acquirer and target capabilities shows that the destruction of value from the acquirer's capabilities exceeds that created from the target's capabilities whenever acquirer overlap is high (quadrants 2 and 4). Consistent with Capron and Mitchell (2009), firms seem to be more efficient at acquiring technologies that are more dissimilar to their own as conflict arises if the firms possess similar technological capabilities. Managers considering acquiring a company with knowledge overlapping much of their own must recognize that such acquisitions are not only unlikely to generate value, they may reduce the value of the acquirer's preexisting knowledge.
Several alternative explanations merit careful consideration. First, as demonstrated by Google's recent acquisition of Motorola Mobility, an acquirer's interest may be less in using the target's technology directly and more in having it available for licensing out or using in litigation-related negotiations (Boulton, 2011). We suspect that few of the small target firms in our sample offer a rich enough IP portfolio to merit purchasing the entire company merely to out-license the technology. To the degree that it does occur, however, the value of the target's IP for licensing (litigation-related or otherwise) would be determined by its value to potential licensees, rather than the acquirer, meaning that its value should be minimally affected by the overlap between the acquirer and target knowledge bases and thus not interfere with the relationships we hypothesize. Second, while our sample was chosen to limit the impact of factors beyond knowledge recombination, other assets of the target, including geographic reach, alliance networks, etc., can create value. Lastly, investors may be concerned if a low target overlap acquisition seems to signal that a firm is moving away from its core capabilities. Benner (2010) found that investors often take this as a negative signal, which would be an alternative explanation for Hypothesis 1a. However, it would work against finding significance for the competing Hypothesis 1b, which was consistently supported.
Future research on the microfoundations of the relationship between overlap, capabilities, and acquisition performance would help address the primary limitation of our study. While we were able to build upon a large theoretical literature and rich empirical findings in developing our hypotheses, we did not directly observe the underlying mechanisms. In particular, it would be helpful to see how the levels of acquirer and target overlap influence the probability of employees departing from the target and, especially, the acquiring firm (cf. Karim, 2006). The same observation could be undertaken for asset divestitures at different levels of acquirer and target overlap. Future research could also generalize our findings to knowledge recombination in acquisitions more generally, although that would require controlling for motives such as market power and cost synergies that may be more prominent.
In conclusion, we have shown that technological overlap in an acquisition is better thought of as two technological overlaps, one describing the knowledge set of the target firm and the other the knowledge set of the acquiring firm. The two overlaps have distinct, but interrelated, effects on the degree to which the acquirer creates (or destroys) value from its own technological capabilities and those of the target. Our findings suggest that the two overlaps drive value creation or destruction through multiple causal mechanisms: the creation of absorptive capacity, knowledge redundancy, and the generation of post-acquisition disruption and conflict. These mechanisms have different theoretical implications and require different managerial responses, which this study has taken the initial steps to explore. We hope others will use this study as a foundation to expand on that exploration.