Make or Buy: The Moderating Effect of R&D-intensity on the Relationship of Performance Feedback and Acquisition behavior.
Student: Rick Joni van der Aar / Student No. 10873791 MSc. Business Administration, Strategy track
University of Amsterdam, Faculty of Economics and Business
Supervisor: MSc. B. Silveira Barbosa Correia Lima
University of Amsterdam, Amsterdam Business School
Statement of originality
This document is written by student Rick van der Aar who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
TABLE OF CONTENTS
Abstract ... 4
THEORETICAL FRAMEWORK ... 10
Behavioral Theory of the Firm ... 11
Performance Feedback and Acquisition Behavior ... 13
Risk ... 16
R&D-intensity ... 19
R&D Intensity vs. Acquisition Behavior ... 20
Differences between aspiration types ... 23
METHODS ... 26
Dependent variable ... 28
Independent variables ... 28
Moderating variable ... 29
Control variables ... 30
Proposed Statistical Models of Regression ... 32
RESULTS ... 33
Descriptives and correlation matrix ... 33
Regression ... 36 DISCUSSION ... 41 Major findings ... 41 Limitations ... 45 Future research ... 46 CONCLUSION ... 47 References ... 48
Abstract
The behavioral theory of the firm states that firms increase their search activity to an alternative strategy, when performance (further) falls below aspirations. To overcome this performance gap, firms become less risk-averse which shapes organizational behavior. This study suggests that firms differ in their response to this performance feedback, with their risk-taking behavior being shaped by a focus on internal development (make) or external acquisitions (buy). A study of acquisition behavior in US manufacturing firms (N= 58,855) using panel data collected between 1980-2014, indicates that R&D expenditures negatively moderate the relationship between performance feedback and a firm’s acquisition behavior. Furthermore, the results indicate that firms that perform below aspiration levels are more likely to attend to historical aspirations when assessing an acquisition decision, whereas firms that perform above aspirations levels are more likely to attend to social aspirations. This paper supplements the behavioral theory of the firm in acquisition contexts, by extending performance feedback theory through analyzing the
moderating effect of R&D-intensity. It further explores the notion that alternative kinds of risk-taking behavior interact with each other and are interdependent, thereby providing a better insight in how performance feedback shapes risk-taking behavior.
Keywords: Performance feedback, behavioral theory of the firm, acquisition behavior,
In recent years, the extent of the research on performance feedback has grown considerably (e.g., Audia & Greve, 2006; Iyer & Miller, 2008; Jordan & Audia, 2012;
Kacperczyk, Beckman, & Moliterno, 2015; Kim, Finkelstein, & Haleblian, 2015). Performance feedback occurs when firms set certain aspiration levels to reflect the firm’s organizational goals and utilize those aspirations as a benchmark for assessing current performance (Cyert & March, 1963; Gavetti, Greve, Levinthal, & Ocasio, 2012). These aspiration levels enable the
interpretation of a firm’s prior performance and can subsequently influence organizational decision-making by modifying managerial risk preferences and search behavior (Cyert & March, 1963; Kim, Finkelstein & Halebian, 2015). The literature has focused on two types of aspiration levels that stem from different sources of performance feedback: social and historical aspirations. The latter is based on the firm’s own past performance, whereas social aspirations emerge from a reference group of competing firms (Greve, 2003a).
According to The Behavioral Theory of the Firm by Cyert & March (1963), firms that fail to meet their aspirations, whether social or historical, engage in problemistic search, the pursuit of identifying alternatives to current activities that resolve performance shortfalls relative to their aspirations (Iyer & Miller, 2008). This can often lead to risky organizational change, as decision-makers feel that they are at a loss and become less risk-averse in devising a solution to improve the status quo (Greve, 2003). In contrast, performance above the aspiration level is seen as the success of the currently applied strategy, meaning there is little incentive for a firm to change its current behavior and to engage in problemistic search (Bromiley, Miller, & Rau, 2001; Greve, 2003). The behavioral theory of the firm has had a tremendous influence on organizational
theory and strategic management. Its concepts lie at the center of many theoretical and empirical studies focused on organizational phenomena (Gavetti et al., 2012).
Decision-makers are key players within an organization who make decisions on its behalf. However, they are constrained in their cognitive capabilities, having bounded rationality or managerial attention (Ocasio, 1997; Greve, 2003a). The consequences of strategic decisions that were made based on performance feedback are vital for decision makers, as their careers depend on these outcomes (Greve, 2011). Scholars have investigated the effect of performance feedback on different types of organizational behavior, such as organizational change, risk-taking, capital expenditures, research and development (R&D) expenditures, and Mergers and Acquisitions (M&A) activity (Bromiley 1991; Greve, 2003b; Audia & Greve, 2006; Chen & Miller, 2006; Iyer & Miller 2008; Kacperczyk, Beckman & Moliterno,2015). Some studies, in accordance with the behavioral theory of the firm, find an increase in organizational change as a response to performance feedback below the aspiration level (Cyert & March, 1963; Halebian et. al.,2006; Greve, 2003b, 2011), whereas others find a decrease in organizational change as a response (Audia & Greve, 2006; Iyer & Miller, 2008; Joseph & Gaba, 2015). These conflicting results have spurred an ongoing debate among scholars, and suggest that firms may respond differently to performance feedback (Iyer & Miller, 2008).
To investigate what causes organizations to respond differently to performance below aspirations, several articles have discussed moderators that may affect the relationship between performance feedback and organizational change (Greve, 2011; Kim, Finkelstein & Halebian, 2016). Furthermore, recent research indicates a wider understanding of the heterogeneity of firms, meaning that firms under different organizational conditions are likely to interpret
performance feedback in particular ways, which leads to distinct organizational risk-taking behaviors (Devers et al., 2007b; Tuggle et al., 2010; Slater, Olson, & Hult, 2006; Kim, Finkelstein & Halebian, 2015). Research from Audia and Greve (2006), for example, has indicated that small firms decrease risk-taking behavior in response to performance below aspirations due to their limited resources and managerial capabilities, although the same effect is not found for large firms as they don’t have the same resource problems.
Two output variables of performance feedback studies that have been studied numerous times are R&D expenditures as a ratio of sales (R&D-intensity) and acquisition likelihood (Greve, 2003; Chen & Miller, 2007; Iyer & Miller, 2008). As both constructs are firm-specific events that can be objectively observed and deemed important by stakeholders, they provide a fertile background for analyzing performance feedback-based organizational behavior. Because further insight in behavioral tendencies of firms in explaining strategic outcomes is highly relevant for scholars and managers alike, this paper further builds on what types of risk-taking behavior as a response to performance feedback are displayed. Particularly, I focus on the moderating effect of a specific organizational condition which is an alternative to acquisionts, in this case R&D-intensity.
Innovations can potentially transform firms and industries, but they are also laden with risk (Greve, 2003). In order to respond to threats and opportunities caused by environmental changes, managers allocate resources to the R&D department (Cohen and Levinthal, 1989). Several authors have elaborated on the relationship between performance feedback and R&D-intensity as a means of improving organization performance with respect to aspiration levels and relationships found in concordance with the problemistic search argument. This means that
R&D-intensity increases when performance falls below aspirations (Greve, 2003; Chen &
Miller, 2007). Contrarily, similar research on the relationship between performance feedback and acquisition behavior by Iyer and Miller (2008) surprisingly found results contradicting the
problemistic search argument, as their results indicate that for underperforming firms relative to their aspiration level, the acquisition count declines as performance falls further. I argue that this contradictory result may be caused by the moderating effect of R&D-intensity on acquisition behavior.
As stated before, both concepts have been studied separately as a consequence of performance feedback and subsequent problemistic search, but not to my knowledge with performance feedback studies in which both variables interact with each other. Iyer and Miller (2008) included R&D-intensity as a control variable in their study on the effect of performance feedback on acquisitions, but did not find it to have a significant effect on acquisition behavior. In contrast, Blonigen and Taylor (2000) found a direct and strong inverse relationship between R&D-intensity and acquisition behavior in US electronic and electrical equipment firms.
The gap in the literature on R&D-intensity as a moderating variable in performance feedback studies is remarkable, as firms may pursue different strategies for growth and survival due to a distinct risk-appraisal and are often presented with make-or-buy decisions: develop internally or acquire externally (Parmigiani, 2006). Make-or-buy decisions occur frequently within firms and are a central theme in transaction cost economics (Williamson, 1975; Perry, 1989; Grant, 1996). While a firm can focus on R&D and M&A simultaneously, both activities are often dependent on each other as they require significant resources to procure (Bigelow & Argyres, 2008). Therefore, a negative correlation between R&D-intensity and acquisition behavior may exist because firms choose between an internal growth strategy with relatively
high R&D-intensity as an alternative risk-taking behavior versus an external growth strategy through acquisitions (Blonigen & Taylor, 2000).
Firms that have high R&D-intensity have a higher probability of generating technical and innovative assets, therefore leading to the prospect of R&D-intensity being positively correlated with acquisition activity. This contradicts the empirical findings of Blonigen & Taylor (2000), who found a strong inverse relation between R&D-intensity and acquisition behavior in a sample of US electronic equipment firms because both concepts act as a substitute rather than
complement eachother. In other words, relatively low R&D-intense firms are more likely to partake in the acquisition market. If R&D-intensity negatively moderates the effect of
performance feedback on acquisition likelihood, it may be that the contradictory results of the relationship between firms performing below aspiration levels and acquisition likelihood as found in Iyer & Miller (2008) can be addressed in more detail. Therefore, the research question is as follows:
What is the moderating effect of a firm’s R&D-intensity on the relationship between performance below aspiration and acquisition behavior?
This study makes two main contributions to the existing performance feedback literature. First, it supplements the behavioral theory of the firm and acquisition behavior by extending previous work on the subject through analyzing the moderating effect of R&D-intensity on the relationship between performance feedback and acquisition likelihood. Therefore, it contributes to a finer understanding of the influences of organizational characteristics on the relationship between performance feedback and organizational risk-taking behavior. Additionally, it explains
how R&D-intensity influences the preferred method of aspiration used by decision-makers when benchmarking, thereby deepening the knowledge about the behavioral implications of the
different aspiration dimensions.
This paper is structured as follows. First, I build a theoretical framework that describes key concepts in performance feedback theory and R&D-intensity. I discuss different perspectives that act as a basis for the research design that I elaborate on in the methods section. The methods section further includes an overview of the sample and describes how data was collected, and which variables are relevant for this empirical study. In the results section, I describe the correlations and the outcomes of the proposed regression analyses. Thereafter, I discuss these results in the discussion section, supplementing them with limitations and suggestions for future research. Finally, the thesis concludes with an overview of the study and a presentation of the most important findings.
THEORETICAL FRAMEWORK
The following section discusses developments in and the current state of performance feedback theory and details the influences of organizational characteristics that may moderate its relationship with strategic change within firms. I discuss the limitations and contradictory nature of the existing literature, which leads to the formulation of hypotheses. Furthermore, I present an overview of relevant risk theories and an overview of make-or-buy literature that reflects on a decision-maker’s choice between a focus on internal development through R&D to boost performance or an acquisition strategy (Blonigen & Taylor, 2000).
Behavioral Theory of the Firm
Performance feedback theory stems from a combination of themes drawn from the
behavioral theory of the firm (Cyert & March, 1963), connected to the “Carnegie School,” which focused on researching organizational behavior in the 1950s and 1960s (Greve, 2003a;
Haleblian, Kim & Rajagopalan, 2006). In the Carnegie school paradigm, the concept of rational choice and perfect information as assumed by neo-classical economists is discarded, as they argue that decision-makers are cognitively limited in their capabilities to act rationally (Gavetti et al., 2012). March and Simon (1958) elaborate on the limited rationality of decision makers within firms, explaining that limited information, attention deficits, and processing ability or lack thereof cause the decision-maker to be unable to maximize performance outputs. Rather than maximizing results, decision-makers follow the cognitive heuristic of satisficing (a combination of “satisfy” and “suffice”). This implies that they set a performance goal that they then try to attain by sequentially searching for and evaluating alternatives for change until a certain acceptability threshold of performance is met (Greve, 2003). Cyert and March (1963) identify two types of search: problemistic search and slack search. Problemistic search is the effort of identifying strategic alternatives to current activities that may deal with performance shortfalls. As it is derived from a problem (the mismatch between performance and aspirations), there is no need to search further for an alternative if the performance goal is met. Slack search takes place when organizations possess excess resources that make it possible to experiment, which may result in identifying and pursuing new opportunities (Levinthal & March, 1981; Iyer & Miller, 2008). Because feedback regarding past performance relative to aspirations can spur search
processes to find and evaluate a suitable solution to address the performance goal, it is considered learning from performance feedback (Greve, 2003).
To conclude, the behavioral theory of the firm has had enormous impact on organizational theory, becoming foundational to many theoretical works on performance feedback, strategic change, the application of search, managerial attention and rationality limitations, and the satisficing nature of decision-makers (Gavetti et al., 2012; Ocasio, 1997; Greve, 2003b).
Performance feedback
Decision-makers attempt to address the constraint of bounded rationality by learning from performance feedback (Greve, 2003; Jordan & Audia, 2012). Based on the behavioral theory of the firm, Levitt and March (1988) define organizational learning as the integration of the organization’s perception of prior experience of the firm itself or other competing firms into the routines that guide the organization’s behavior. How they interpret and integrate these past experiences depends on the performance goals, or in other words, the aspiration levels (Levitt & March, 1988). Organizations set an aspiration level and use it as a benchmark to evaluate their performance, assessing whether their performance is considered a success or failure (Cyert & March, 1963; Shinkle, 2012). They are derived from comparisons of two sorts of reference points: historical performance and social performance. In historical aspirations, the
organization’s (recent) past performance acts as reference point whereas the social aspiration level is stems from a comparison of the performance of a peer group of firms (Cyert & March, 1963; Levinthal & March, 1981). Aspiration levels can be seen as the smallest outcome that would be deemed satisfactory by the decision maker (Greve, 2003). Differences between the
organization’s observed performance and aspirations is termed an attainment discrepancy, which determines whether the firm will change its organizational behavior (Iyer & Miller, 2008).
Literature on performance feedback provides ample empirical evidence that firms that perform below aspirations have the tendency to become more prone to risk-taking behavior, such as increased spending on R&D or commencing an acquisition strategy, due to a negative
attainment discrepancy which induces a problemistic search (Greve, 1998; Greve, 2003; Greve, 2010; Iyer & Miller, 2008; Nickel & Rodriguez, 2002). This increased risk-taking is caused by a manager’s desire to overcome performance gaps as his career of often dependent on the firm’s performance. In contrast, organizations performing above their aspirations are less inclined to change, as managers feel that they are ‘on the right track’ and get strengthened in their beliefs to persist the current strategy. (Greve, 2003). By linking change to organizational issues, as
expressed through performance below aspirations, performance feedback research has provided an empirical foundation for models of organizational learning, search, and adaptation that play a significant role in the behavioral theory of the firm, as it explains why firms take more risk during performance shortfalls (Gavetti et al., 2012).
Performance Feedback and Acquisition Behavior
Most studies in the performance feedback domain have focused on performance measures such as return on assets (ROA), which are considered top-level organizational
priorities (Gavetti et al., 2012). Therefore, these studies often employ research designs which are better suited for predicting actions that top-level decision-makers make (Gavetti et al., 2012), such as mergers and acquisitions (Finkelstein, Halebian, & Kim, 2015; Haleblian, Kim, &
Rajagopalan, 2006; Iyer & Miller, 2008), diversification (McDonald & Westphal, 2003), market position change (Park, 2007), and growth (Audia & Greve, 2006; Desai, 2008; Greve, 2008).
Mergers and acquisitions are a complex undertaking and fail often (Zollo & Singh, 2004). Studies by Porter (1989) and Young (1981) suggest that acquisitions have a high failure rate – about half of all acquisitions are appraised as being unsatisfactory by managers of acquiring firms.
In spite of a poor track record in this area and the risks, organizations nonetheless continue to consider acquisitions as a genuine strategic option (Kim, Haleblian, & Finkelstein, 2011). As acquisitions can constitute substantial performance turnarounds for firms, Iyer & Miller (2008) hypothesized that, following problemistic search, when a firm performs below aspirations, the probability of an acquisition increases. However, contrary to Iyer and Miller’s (2008) expectation, they found that for underperforming firms relative to their aspiration level, a further decrease of performance actually leads to a decrease in the likelihood of an acquisition. This seemingly contradicts the theory of problemistic search that predicts an increase due to decision-makers becoming more risk-seeking (Greve, 2003), or implies that acquisition behavior is not the dominant form of risk-taking behavior for low performing firms.
There has been more scientific literature that challenges the basic idea of the relationship between underperforming firms and increased strategic change, arguing instead that
organizations are likely to process information on performance feedback differently, thereby bringing about distinctive organizational-dependent risk-taking behavior (Devers et al., 2007b; Greve, 2011; Lim & McCann, 2014). Audia and Greve (2006) tried to reconcile these two streams of literature, and found that underperforming small firms reduce risk-taking behavior, whereas larger firms increase risk-taking behavior due to having a larger pool of resources,
which buffers them from the threat of failure or bankruptcy. The rationale behind this is that smaller firms are constrained by their limited financial and managerial resources that could help overcome the problems that threaten their survival (Mitchell et al, 1994). This extends the performance feedback theory by theorizing that risk-appraisal is dependent on firm
characteristics, and that firms may shift their focus of risk-taking behavior accordingly. Also, firms that reach their “survival-point,” at which performance has dropped so much that their survival is threatened, shift their focus from risk-taking behaviors induced by problemistic search to survival, meaning they become more risk-averse (Audia & Greve, 2006). This implies that there are organizational characteristics (e.g. firm size, pool of resources) that change the way companies react to performance feedback due to a shift of focus (Greve, 2011).
In contradiction, Finkelstein, Halebian, and Kim (2015) argue that while problemistic search in acquisition contexts occurs, overall firm performance measures such as ROA – as used by Iyer and Miller (2008) – are “too general and far removed from the vicinity of the problem in which problemistic search occurs.” Iyer and Miller (2008) themselves give a possible
explanation for their contradictory results by arguing that below-aspiration performers may engage in a local search towards enhancing performance rather than turning to acquisition
behavior. Iyer and Miller (2008) invite other researchers to investigate decision-makers’ focus on risk-taking within firms and when this shifts to another type of risk-taking behavior other than acquisitions. Consequently, other researchers have proposed a variety of moderators that may affect the relationship between performance feedback and organizational change (Greve, 2011). Later in this paper, I introduce R&D-intensity as a moderating variable as an alternative form of risk-taking behavior, but before that I propose two hypotheses to test the contradicting results
found by researchers regarding performance feedback for underperforming firms in acquisition contexts.
Hypothesis 1a: The relationship between performance below aspiration on acquisition count is
negative, such that a further decrease relative to aspirations leads to an increase in acquisition count.
On the other hand, contradictory results such as those found by Iyer and Miller (2008), also provide ample evidence that the perceived relationships derived from the problemistic search argument are not always present due to a shift of focused caused by different risk-appraisals. Thus, to shed more light on this argument, I alternatively propose:
Hypothesis 1b: The relationship between performance below aspiration on acquisition count is
positive, such that a further decrease relative to aspirations leads to a decrease in acquisition count.
Risk
Firms need both a motive and the necessary resources to engage in an acquisition (Iyer & Miller, 2008). As mentioned, these motives are not solely guided by performance feedback, but are dependent on firm-specific characteristics (Greve, 2011) and managerial attention (Ocasio, 1997). An example of these characteristics that is central to performance feedback theory, is the way risk is interpreted (Gavetti et al., 2012). As stated before, Iyer and Miller (2008) found evidence for a shift in the focus of managerial attention when firms perform below historical
aspirations and near to their survival-point. A possible reason for finding such a relationship is further built upon by Gavett et al. (2012) who discuss the moderating effect of risk in
performance feedback. They argue that firms that are financially at risk instead become rigid when performing below aspirations. This is called the threat-rigidity hypothesis, which suggests that organizations that are focused on survival are less likely to commit themselves to risk-taking behaviors (Staw, Sandelands, & Dutton, 1981).
Firm size also has an effect on how risk is interpreted. Audia and Greve found that capital investment is more likely to be conducted by large organizations below aspiration level than by small firms (Audia & Greve, 2006; Greve, 2011). Extrapolating this to an acquisition context, it may that acquisitions are more likely to be conducted by large organizations than by smaller ones. Furthermore, large organizations are more likely to acquire qualitatively different assets than small ones (Greve, 2011). Therefore, searching in proximity to current practices is more likely for firms that are small or otherwise limited in their resources, as this is deemed less risky (Gavetti et al., 2012). Organizations that are not threatened in their survival but
experiencing poor performance also often resort to acquisitions to supercharge their growth (Slater, 1984: 96). Empirical work from Sudarsanam and Lai (2001) has shown that
organizations that are in the process of recovering from distress more often choose acquisition to lead them (further) out of trouble, whereas non-recovering firms are inclined to be more
internally focused on operational and strategic problem-solving. To summarize, what is interpreted by one firm as rigorous change with high risk involved may be interpreted as a standardized and less risky choice by another firm depending on a combination of contextual factors (Greve, 2011; Ocasio 1997).
Apart from firms that focus on an M&A strategy, firms can alternatively invest more in R&D to induce change derived from development from within the organization. However, for an organization that is very focused on R&D, deciding to invest less in innovations is also a form of organizational change, leading to the argument that risk is not the necessarily the same as change (Kacperczyk et al, 2015). Kacperczyk et al. (2015) define risk as “the variation in the distribution of possible outcomes and uncertainty associated with gains and losses”, whereas organizational change is not necessarily uncertain or varied in outcome. They theorized that risk-taking
behavior and organizational change are implemented by managers under different conditions, such that change is driven by organizational problems, whereas risk is driven by individual concerns. Following the behavioral theory of the firm, when performance feedback indicates a discrepancy between aspirations and performance, managers are likely to induce a change
process through problemistic search that will satisfice the performance goals. In contrast, risk is a response to performance feedback indicating an individual concern that is triggered by loss aversion (Kacperczyk et al., 2015). Loss aversion is the tendency of people to prefer avoiding losses to obtaining equivalent gains. Managers respond to this loss aversion by taking greater risk within the company (Kacperczyk et al., 2015).
Many authors in the performance feedback have failed to distinguish change to risk and vice versa. Therefore, while a major outcome variable for strategic change in performance feedback studies (Greve, 2003; Chen & Miller, 2007; Vissa, Greve & Chen, 2010, Lim & McCann, 2014), R&D expenditure is not always induced from performance feedback but could be an persistence of an existing strategy. Acquisitions, on the other hand, constitute an
organizational change (Kacperczyk et al., 2015) and are often risky due to their high failure rate and uncertain outcome. Theories of risk broaden the theory of performance feedback specifically
in a way that offers an explanation for firms’ different reactions to performance feedback based on differences in managerial attention to risk. In the following section, I discuss how R&D expenditure and acquisition behavior are related and how the former is based on differences in risk-appraisal.
R&D-intensity
Problemistic search starts with looking for local solutions, and then broadens “to include progressively more distant possibilities if initial efforts fail to uncover a ‘sat-isficing’ [sic] alternative” (Cyert & March, 1963; Levinthal, 1997). Local search is defined as searching alternatives that lie within the vicinity of the problem. Following this argument, organizations initially search by examining operational and strategic actions within the firm before considering changes that include acquiring external resources. Both R&D expenditures and acquisitions have been researched comprehensively in the performance feedback domain because these outcomes are highly consequential and generally dependent on top managers’ financial commitment (Gavetti et al, 2012). Similar to the case of acquisitions, managers in underperforming firms can invest in R&D in an attempt to ensure performance turnarounds and accelerate growth by exploring long-term creation of genuinely novel innovations (Antonelli, 1989; Hundley, Jacobson & Park, 1996; Greve, 2003). However, R&D expenditure represents significant
financial costs in the short term and uncertain outcomes with no guaranteed pay-off whatsoever, and could even include losses (David et al., 2001, in Lim & McCann, 2014).
R&D is an operational cost, meaning decreasing it would directly entail an improvement of company performance (ROA). Nonetheless, top decision-makers are more likely to see R&D as a search activity that can provide solutions for organizational challenges, and therefore often increase R&D expenditures to bolster innovations.
R&D-intensity refers to the ratio between R&D expenditure and has been used several times in performance feedback studies as a proxy for strategic change (Greve, 2003a; Chen & Miller, 2007; Vissa, Greve & Chen, 2010; Lim, & McCann, 2014). Literature on the effects of R&D-intensity on firm performance has indicated mixed results. For example, Falk (2010) found that initial R&D-intensity had a positive and significant impact on both employment and sales growth in the subsequent two years for Austrian firms between 1995-2006, whereas Lin, Lee, and Hung (2006) found no significant relationship between R&D-intensity and firm performance in 258 US-based technology firms.
To conclude, R&D-intensity has been a focal topic in performance feedback studies as a proxy for strategic change. It is often costly with uncertain outcomes, but is nonetheless
considered by managers as a sensible, and in many cases necessary, search activity that can provide solutions for organizational challenges. In the following section, I discuss the relationship between R&D-intensity and acquisition behavior further.
R&D Intensity vs. Acquisition Behavior
Firms achieve new innovations either through internal R&D or through acquisitions, often labeled a “make-or-buy” decision (Xue, 2007). The make-or-buy decision originates from transaction cost theory, first proposed by Coase (1937) but further developed notably by
Williamson (1985), which refers to the nature of the firm concerning which activities should be conducted from within the firm and which should be acquired externally. Xue (2007) builds on this foundation by describing how firms choose between R&D (make) or acquiring (buy) in high-tech companies. While this seemingly implies that companies choose either a focus on R&D or an acquisition strategy, both can be utilized simultaneously, depending on specific firm and market characteristics (Parmigiani, 2007).
Scholars have found that both concepts are, however, intertwined. Several studies have found that acquisitions can alter firms’ incentives to innovate internally (Gans & Stern, 2000; Katz & Shapiro, 1986; Salant, 1984). Blonigen and Taylor (2000) found a strong negative correlation between R&D-intensity and acquisition activity in US electronic and electrical equipment firms. Or stated alternatively, relatively low R&D intensity firms are more likely to participate in the acquisition market. This direct and strong inverse relationship is best illustrated with an example: a firm with a 5% higher R&D-intensity has an approximately 26% lower yearly acquisition rate. Hall (1990) studied the relationship between M&A and R&D-intensity as an ex post relationship, hypothesizing that R&D efforts would decline after an acquisition, rather than its potential role as a factor in acquisition decisions by firms. He found proof, however, for an ex ante relationship of acquiring firms tending to have lower R&D expenditures relative to the rest of their industry, theorizing that those firms have chosen an acquisition strategy focus. Furthermore, Friedman et al. (1979) studied the relationship between R&D-intensity and joint venture activity (as opposed to acquisition behavior) and found that the greater the involvement of firms in joint venture activities, the lower the R&D expenditures, theorizing that joint ventures act as a substitute for internal R&D activity.
In contrast, there have been also several studies that suggest complementarity between R&D-intensity and external acquisitions (Cassiman and Veugelers, 2006). Although transaction cost theory suggests that the availability of external knowledge may substitute for internal R&D investment (Williamson, 1985, Pisano, 1990), Cassiman and Veugelers (2006) state that “both causal evidence and more rigorous empirical research suggest the existence of complementarity between in-house R&D and external knowledge.” Central to this is the extent to which the organization is capable of effectively “absorbing” external innovations (Cohen & Levinthal, 1990, Allen 1986).
The way organizational behavior is shaped, whether “make” or “buy,” is dependent on managerial attention. Managerial attention is in turn dependent on formal and informal firm characteristics, and may lead to firm-specific risk appraisals. Xue (2007), for example, found that executive compensation plans are a significant predictor of how the make-or-buy decision is made. As theorized by Kacperczyk et al. (2015), risk is a response to performance feedback indicating an individual concern that is triggered by loss aversion. Xue’s study (2007) found that decision-makers whose pay heavily depends on accounting performance measures (e.g. net income) will tend to acquire (buy) rather than by R&D (make) to avoid high uncertainty and the direct negative effect it has on those measures. On the other hand, decision-makers who hold a myriad of stock options or stocks are more likely to invest in R&D because these mitigate disadvantages due to managerial risk-aversion and required accounting treatment (Xue, 2007).
To conclude, a negative correlation between R&D-intensity and acquisition activity may manifest because firms choose between an internal growth strategy with relatively high R&D-intensity versus an external growth strategy with acquisitions. But results from the literature are
mixed, sometimes implying complementarity and sometimes dependence on managerial attention and risk-appraisal. These discrepancies, however, can explain how firms in the same industry pursue different strategies for change as a response to performance feedback. Following the problemistic search argument, local solutions like R&D are likely considered before external solutions such as acquisitions. Therefore, I reason that as a response to performance feedback, R&D-intensity is negatively moderated to acquisition behavior due to firms choosing between a search for internally induced change or external search.
Hypothesis 2: R&D-intensity negatively moderates the effect of performance below aspiration on
acquisition count.
Differences between aspiration types
Performance feedback studies have resulted in significant evidence that organizational change increases when performance is below aspirations and slows when it rises above (Audia et al., 2000; Greve, 1998; Greve, 2003; Greve 2010). Aspirations are divided into a historical level and a social level, but whether and how they generate different influences on firm behavior was not studied until recently, by Finkelstein, Halebian, and Kim (2015). They were surprised by this omission, given that most performance feedback tested both types of aspirations separately and has often yielded inconsistent empirical findings (Audia & Greve, 2006; Greve, 2003b; Iyer & Miller, 2008; Shipilov, Li, & Greve, 2011). Several studies have found more significant effects for historical aspiration levels than for social aspirations (e.g., Audia & Greve, 2006; Greve, 2003b; Shipilov, Li, & Greve, 2011), and vice versa (e.g., Harris & Bromiley, 2007).
Finkelstein, Halebian, and Kim (2015) posit that historical and social aspiration levels may induce contrasting interpretations as they emerge from different performance information sources and are filtered through dissimilar cognitive organizational processes, and that they effect firms’ behavior and choices accordingly.
Since historical aspiration is derived from past performance, it closely mirrors the managerial capabilities and resources an organization brings to an acquisition (Finkelstein, Halebian & Kim, 2015). These factors make the historical aspiration level a relatively credible indicator of how well an organization could perform, given its resources and capabilities (Greve, 2003a: 42). Social aspiration levels, on the other hand, help decision-makers assess how well they
should perform, considering that stakeholders will expect the organization to perform at least on
par with other competing firms in the industry. Following this logic, recent studies on aspirations imply that decision-makers first focus on social aspirations, as this implies the baseline performance level (“how well they should perform”) before they attend to other performance benchmarks (Audia & Brion, 2007; Washburn & Bromiley, 2012).
In general, historical performance is subject to deeper inspection by managers, as they have access to private knowledge that is located within the organization and may use such knowledge to interpret historical performance and identify factors contributing to the performance outcomes (Menon & Pfeffer, 2003). Social aspiration levels, on the other hand, are often ambiguous, and it is difficult to identify the underlying factors that contribute to the outcome observed (Baum & Ingram, 2002).
To conclude, performance feedback based on different kinds of aspirations may result in different firm behavior, as it is derived from different information sources and organizational
processes. I expect that firms with relatively low R&D-intensity are predisposed to rely on historical aspirations than social aspirations. R&D expenditures are uncertain and it could take firms years to see pay-offs realized. Therefore, firms not used to these uncertainties consider less ambiguous performance measures, such as historical aspiration level, first, as they constitute managerial capabilities and resources more closely.
Hypothesis 3a: When acquiring, decision-makers in firms with relatively low R&D-intensity and
performing below aspirations are more likely to attend to historical aspirations than social aspirations.
On the other hand, I expect firms with relatively high R&D-intensity are more inclined to rely on social aspirations than historical aspirations. These organizations are used to the
uncertain outcomes of R&D expenditures and therefore less nervous regarding ambiguous
performance measures, such as social aspirations. Furthermore, due to this ambiguity, these firms already have a more external focus on peer performance, and therefore a better grasp of how they
should perform.
Hypothesis 3b: When acquiring, decision makers in firms with relatively high R&D-intensity and
performing below aspirations are more likely to attend to social aspirations, than historical aspirations.
METHODS
In this section, I describe the sample and the data collection procedure. I explain and justify the variables and provide some general descriptives of the sample. Thereafter, I depict the statistical model and methodological approach.
Sample
The dataset that I used for this research is constructed from data collected from the Securities Data Corporation (SDC) “Platinum” database and merged with corresponding annual financial and industry data from the CRSP-Compustat database. In concordance with other researchers on performance feedback and acquisition behavior, I restricted these analyses to US manufacturing industries (SIC codes 2000 to 3999) as this enables a more reliable comparison of the results by preventing confounding results owing to major differences in industries’ activities (Iyer & Miller, 2008; Chen & Miller, 2007).
This empirical research solely focuses on acquisition behavior as a strategic decision, not for any other reasons such as unrelated corporate trading. Therefore, I have only included SDC data on acquisitions where a change of control in ownership has occurred at the ultimate parent level (meaning that the acquiring company has less than 50% of the shares of the target company before acquiring), and have excluded any minority stake purchases (meaning that the acquiring firm bought at least 50% of shares of the target company during the acquisition). Furthermore, spinoffs, recapitalizations, self-tenders, exchange offers, repurchases, and privatizations were
excluded from the sample as they are not relevant transactions in which performance feedback mechanisms occur.
Twenty-eight point eight percent of the firm-year observations were initially missing data on R&D expenses, which is caused by firms not required by the SEC to separate R&D
expenditure from operational costs if they are less than 10%. Following Iyer and Miller (2008), I exclude these missing values from the data analysis. Chen and Miller (2007) however, when confronted with the same missing data in a similar sample and research design, replaced the missing values with zero on the assumption that these organizations invested small amounts in R&D (less than 10% of operational costs). When running the same analyses with missing values excluded from the dataset, they obtained essentially the same results.
Furthermore, I restrict my sample to firms with an R&D-intensity less than or equal to 1, as my reasoning solely applies to manufacturing firms committed to ongoing production and sales activities and not R&D specialists.
After these omissions from the sample, 58,855 firm-year observations remain, including 10,309 acquisitions by 5,766 different companies, using data from 1980 to 2014 in fiscal years. The four largest two-digit industry divisions are Chemicals & Allied Products (SIC 2800) accounting for 21.2% of all firms in the sample, Electronic & Other Electrical Equipment (SIC 3600) with 21.0%, Instruments & Related Products (SIC 3800) with 17.7% of the firms in the sample, and Industrial Machinery & Equipment (SIC 3500) with 16.9% of the firms in the sample.
Dependent variable
Acquisition behavior. The dependent variable that is analyzed in this study is acquisition behavior, or specifically, acquisition count. Acquisition count is defined as the frequency of acquisitions performed in one fiscal year. The mean of acquisition count lies at 0.28 acquisitions per fiscal year, and the standard deviation is 0.82. The frequency of acquisitions is a count variable in which observations can only take non-negative integer values, and refers to the
number of times a specified event occurs. For a further overview of acquisition count frequencies within the sample, please refer to Table 1.
Independent variables
Return on assets. In this study, I use return on assets (ROA) as a proxy for firm
performance which is computed by dividing net income with total assets. It is the main measure of an organization’s profitability and it has been applied as such in many performance feedback studies (Audia & Greve, 2006; Bromiley, 1991; 2004; Desai, 2008; Greve, 2003; Iyer & Miller, 2008; Lim & McCann, 2014).
Historical aspirations are measured as the firm’s average performance (ROA) over the past three years.
Social aspirations are measured as the mean performance of firms in the same three-digit SIC industry codes over the period t–2 in a similar way as other performance feedback studies (Chen & Miller, 2007; Iyer & Miller, 2008).
Performance feedback is the difference between the firm’s ROA and the aspiration level, either social or historical. To test acquisition count as a response to performance being higher or lower than aspirations, two variables were created, splitting the firm’s aspiration levels into performance above and performance below aspirations (Greve, 2003b).
Performance above aspirations is the difference between ROA and aspirations, but only for the observations where ROA is higher than the aspiration level, meaning all variable data is greater than zero.
Performance below aspirations, on the other hand, is defined as the difference between ROA and aspirations in which ROA is lower than the aspiration level, meaning all variable data is less than zero (Greve, 2003b; Chen & Miller, 2007).
Moderating variable
R&D-intensity. Some of the literature about R&D-intensity uses different definitions, which is defined as the ratio of R&D expenditure to sales, or the ratio of R&D expenditure to total assets. Earlier research defines this as the latter (Hall, 1987; Blonigen & Taylor, 2000) whereas more recent work focuses on the R&D ratio with sales (Greve, 2003; Chen & Miller, 2006; Iyer & Miller, 2008). Although research suggests that both constructs lead to essentially identical results (Blonigen & Taylor, 2000), I use the ratio of R&D to assets.
Control variables
Firm size. As found by Audia and Greve (2006), firm size moderates the relationship between performance feedback and risk-taking behavior. Therefore, I control for firm size in a similar manner as previous scientific work by logging the number of employees (Audia & Greve, 2006; Greve, 2003b; Gaba & Joseph, 2013).
Potential slack. Slack is possession of excess resources that allow for experimentation. Potential slack refers to the ratio of debt to equity (Iyer & Miller, 2008) but can be visualized as the potential future resources that can be generated by the organization from outside sources by, for example, issuing stock or bonds or taking out a loan.
Unabsorbed slack is assets divided by current liabilities, alternatively known as the current ratio (Iyer & Miller (2008), meaning the resources are readily available. Iyer and Miller (2008) found that potential and unabsorbed slack had a significant effect on acquisition
likelihood. Therefore, controlling for slack can mitigate for the different resource levels across firms.
Table 1
Acquisition events per year
Acquisition count
Fiscal Year 0 1 2 3 4 5 6 >6 Total %
1979 196 196 0.3 1980 1,252 12 2 1,266 2.2 1981 1,279 53 6 1 1,339 2.3 1982 1,275 62 7 1 1,345 2.3 1983 1,421 84 11 2 1,518 2.6 1984 1,437 112 12 6 1 1,568 2.7 1985 1,457 91 22 10 2 1,582 2.7 1986 1,441 139 26 17 3 1 1 1 1,629 2.8 1987 1,491 126 26 11 4 2 1 1 1,662 2.8 1988 1,423 154 27 12 1 2 1 1,620 2.8 1989 1,372 167 45 8 4 2 1 1,599 2.7 1990 1,330 180 35 12 5 6 1 1,569 2.7 1991 1,378 184 52 11 5 1 1 1,632 2.8 1992 1,424 231 52 15 5 1 2 1 1,731 2.9 1993 1,529 248 51 22 11 1,861 3.2 1994 1,548 267 72 23 11 2 1 1,924 3.3 1995 1,588 276 95 28 17 4 3 3 2,014 3.4 1996 1,728 299 97 40 16 8 4 6 2,198 3.7 1997 1,699 339 115 30 21 10 3 9 2,226 3.8 1998 1,542 324 102 50 23 13 7 7 2,068 3.5 1999 1,482 267 100 36 28 11 4 8 1,936 3.3 2000 1,524 289 99 39 22 9 8 9 1,999 3.4 2001 1,530 248 74 24 16 6 2 4 1,904 3.2 2002 1,473 215 65 33 2 5 1 2 1,796 3.1 2003 1,406 228 67 23 12 9 6 1,751 3.0 2004 1,392 259 71 45 7 4 2 5 1,785 3.0 2005 1,377 256 62 30 12 6 2 6 1,751 3.0 2006 1,342 235 79 36 10 6 2 10 1,720 2.9 2007 1,312 241 65 27 14 6 4 9 1,678 2.9 2008 1,214 227 66 20 13 2 4 6 1,552 2.6 2009 1,198 196 50 19 6 5 3 2 1,479 2.5 2010 1,178 176 70 22 7 5 2 6 1,466 2.5 2011 1,109 197 69 36 6 3 6 6 1,432 2.4 2012 1,120 189 54 25 10 3 4 7 1,412 2.4 2013 1,157 191 60 18 6 4 6 1,442 2.5 2014 922 184 57 23 10 3 3 3 1,205 2.0 Total 48,546 6,946 1,963 754 311 139 76 120 58,855 100
Proposed Statistical Models of Regression
The sample contains panel data involving measurements over time and a dependent discrete count variable. Therefore, Poisson or negative binomial regression analyses are the preferred choice of statistical model (Gaba & Joseph, 2013; Joseph & Gaba, 2015). In order to choose which technique is most appropriate, I conduct a preliminary analysis with a General Linear Model to check for the Pearson dispersion statistic. This statistic indicates whether data exhibits over-dispersion, meaning the variance in the dataset is higher than one would reasonably expect. However, the Pearson χ2 dispersion statistic (1/df Pearson) indicates a value of 0.72, meaning there is no problem with over-dispersion within the dataset. Therefore, I can perform a Poisson regression analysis, which assumes the mean and variance to be the same.
Subsequently I conducted a Hausman’s specification test to decide whether to employ the within (fixed effects) or random effects estimator (Hausman, 1978). This test resulted in
rejecting the null hypothesis ((8) = 20.85, p < 0.01), meaning a random effects estimator cannot adequately account for firm and time effects in the dataset. Therefore, all regression analyses are executed by using the within (fixed effects) model.
Because I employ a moderating variable in testing hypothesis 2, multicollinearity problems caused by the interaction effect may occur (Lim et al., 2014). Multicollinearity
problems may increase the variance of the coefficients estimates and result in the estimates being very sensitive to minor changes in the model. To prevent this, I mean-centered all independent variables before constructing the interactions. Since this research argues that acquisition behavior is predicted by performance feedback, the independent and control variables are lagged one year relative to the dependent variable (acquisition count). Thus, the independent and control
variables data range from 1979-2013, whereas the dependent variable corresponds to the years 1980-2014. I used the STATA statistical package to perform the within (fixed effects) models.
RESULTS
The following section describes the results of this research. First, I provide descriptives and a correlation matrix to present an overview of the data and the interrelations of the variables. Afterwards, I present the results of the Poisson regression analyses in order to test the hypotheses that were formulated earlier.
Descriptives and correlation matrix
Table 1 provides descriptive statistics and the bivariate correlation analysis of the variables for the dataset. The relationships were assessed by using the Pearson r
correlation coefficient, which is a measure for the degree of linear dependence between two variables. As expected, all the performance feedback measures are significantly correlated with acquisition count, but the correlations are very weak (r < .10). R&D-intensity is very weakly negatively correlated with acquisition count (r = 0.09), moderately with performance below social aspirations (r = 0.41), and very weakly with performance above social aspirations (r = 0.08). Performance below historical aspirations is moderately positively correlated with R&D-intensity (r = 0.30), and performance above historical aspirations is very weakly positively correlated with R&D-intensity (r = 0.12). For the control variables, only firm size shows a (negative) moderate correlation with R&D-intensity (r = 0.40), and some (very) weak
correlations with the performance feedback measures. The slack variables all have very weak, and only in a few cases significant, correlations with the other variables.
Descriptive Statistics and Correlations M SD Minimum Maximum 1 2 3 4 5 6 7 8 9 1. count 0.28 0.82 0.00 25.00 1.00 2. Firm size 6.75 2.24 0.00 13.68 0.24* 1.00 3. Potential slack 0.65 19.34 -782.55 3096.64 0.00 0.01* 1.00 4. Unabsorbed slack 3.95 5.53 0.01 318.82 -0.06* -0.33* -0.01* 1.00 5. R&D-intensity 0.09 0.13 0.00 1.00 -0.09* -0.40* -0.01* 0.13* 1.00 6. Performance below aspirations (historical) -0.07 0.20 -22.57 0.00 0.06* 0.17* 0.00 0.00 0.30* 1.00 7. Performance above aspirations (historical 0.06 0.21 0.00 17.07 -0.03* -0.19* 0.00 0.07* 0.12* 0.10* 1.00 8. Performance below apsirations (social) -0.08 0.25 -22.47 0.00 0.07* 0.27* 0.00 0.01 -0.41* 0.80* -0.03* 1.00* 9. Performance above aspirations (social) 0.12 0.16 0.00 2.95 0.07* 0.04* -0.01 0.05* -0.08* 0.17* 0.09* 0.22* 1.00
In Tables 3 and 4, I present an overview of the Poisson regression results. Table 3 consists of models 1 to 5, whereas Table 4 consists of models 6 to 9. Model 1 is the base model with control variables. I add the independent variables in models 2 and 4, indicating the main effects of the model. Models 3 and 5 introduce the interaction terms between R&D-intensity and the performance feedback measures in order to examine whether a moderating effect is present. For models 6 to 9, I propose a shift-of-focus model, therefore (median) splitting the sample into firms with relatively low R&D-intensity and firms with relatively high R&D-intensity. Models 6 and 8 represent the base model with the control variables. Models 7 and 9 include the
performance feedback measures.
According to the interpretation of earlier performance feedback studies (Chen & Miller, 2007; Iyer & Miller, 2008), a negative regression coefficient in firms that are performing below their aspirations means that the lower past performance drops below the aspiration level, the higher the acquisition count, whereas a positive coefficient implies that the further the past performance falls below their aspirations, the lower the acquisition count.
Hypothesis 1a theorizes that the relationship between performance below aspiration on acquisition count is negative, such that a further decrease relative to aspirations leads to an
increase in acquisition count, whereas hypothesis 1b posits that a further decrease relative to
aspirations leads to a decrease in acquisition count. Looking at models 2 and 4, one sees that for performance below historical aspiration (t = 1.082, p <0.01) and performance below social aspiration (t = 0.606, p <0.01) are both significant, indicating that the further the performance falls below aspirations, the lower the acquisition count. Therefore, this relationship contradicts hypothesis 1b and allows for the acceptance of hypothesis 1b.
Hypothesis 2 implies that R&D-intensity negatively moderates the effect of performance below aspiration on acquisition count. Looking at the full models with interaction effects, one sees that the main effect of R&Dintensity on acquisition count is negative in model 3 (t = -0.944, p <0.01) and model 4 (t = -0.796, p <0.001). Furthermore, model 3 indicates that the interaction between historical performance below aspirations and R&D-intensity is significant (t = -2.099, p <0.01, implying that R&D-intensity negatively moderates the relationship. Model 5 indicates that the interaction between social performance below aspirations and R&D-intensity is also significant (t = -1.341, p <0.01). The negative interaction coefficient implies that the more negative the R&D-intensity coefficient is, the more positive the effect of performance below aspirations becomes on acquisition count. As implied before, the higher the coefficient of performance below aspirations, the lower the acquisition count. Therefore, it can be concluded that the interaction effect will further strengthen the decrease of acquisition count that was already caused by R&D-intensity. Therefore, I conclude that the moderation effect is present and negative as hypothesized, and hypothesis 2 is supported.
One can extend this logic for the interaction between R&D intensity and performance feedback above aspirations, which is significant for social aspiration level (t = 2.627, p <0.01) but not significant for historical aspiration level (t = -0.819, p =0.262). Exemplifying the former, the positive interaction coefficient implies that the more negative the R&D-intensity coefficient is, the lower the effect of performance above social aspirations becomes on acquisition count. Or in other words, the interaction effect will further strengthen the decrease of acquisition count that was already caused by R&D-intensity, concluding that the moderation effect is present and negative.
For hypothesis 3 I have (median) split the sample in relatively low R&D-intensity firms and relatively high R&D-intensity firms. By doing so, I am able to check for a potential shift-of-focus effect. The cut-off point is 0.05, meaning that firms with R&D expenditures higher than 5% of total assets are considered highly R&D-intensive firms.
Hypothesis 3a posits that decision-makers in firms with low R&D-intensity who perform below aspirations are more like to concentrate on historical aspirations than social aspirations. Model 7 indicates that this is indeed the case with performance below historical aspirations being a significant acquisition count (t = 2.323, p <0.01) whereas performance below social aspirations is not (t = 0.367, p = 0.274). Hence, hypothesis 3a is supported.
I further hypothesized that high R&D-intensity firms that perform below aspirations are more likely to attend to social aspirations than historical aspirations (hypothesis 3b). Model 9 indicates that there is no significant relationship between performance below social aspirations and acquisition count for high R&D-intensity firms (t = -0.049, p = 0.781), rather indicating that for these firms, historical aspirations are a significant predictor (t = 0.485, p < 0.01) just as with low R&D-intensity firms. Therefore, no proof for hypothesis 3b is found. Hence, the hypothesis is rejected. It is interesting to remark that for performance above historical aspirations, no
significant relations are found with acquisition count for both low and high R&D-intensity firms. For both low and high R&D-intensity firms, the relationship between performance above social aspirations and acquisition count is significant.
Poisson Regression Model for Acquisition Count (fixed effects)
Historical Aspiration Social Aspiration
Model 1 Model 2 Model 3 Model 4 Model 5
Variables Control SE Main SE Interaction SE Main SE Interaction SE Firm size 0.275** 0.014 0.263** 0.015 0.261** 0.015 0.268** 0.014 0.267** 0.014 Potential slack 0.000 0.000 0.000 0.001 0.000 0.001 0.000 0.000 0.000 0.000 Unabsorbed slack 0.027** 0.003 0.027** 0.001 0.000** 0.004 0.021** 0.003 0.021** 0.003 R&D-intensity -0.741** 0.241 -0.944** 0.243 -0.712** 0.225 -0.796** 0.230 Performance below aspirations 1.082** 0.121 1.306** 0.130 0.606** 0.111 0.771** 0.129 Performance above aspirations 0.190** 0.073 0.179* 0.073 0.617** 0.056 0.735** 0.070 Performance below aspirations x
R&D-intensity -2.099** 0.392 -1.341** 0.287
Performance above aspirations x
R&D-intensity -0.819 0.730 2.627** 0.841
Number of observations 33,847 31,182 31,182 33,744 33,744
Number of acquisitions 2,320 2,146 2,146 2,315 2,315
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Table 4
Poisson Regression Model for Acquisition Count with Low-High R&D-intensity sample (fixed effects)
Low R&D-intensity High R&D-intensity
Model 6 Model 7 Model 8 Model 9
Variables Control SE Main SE Control SE Main SE
Firm size 0.209** 0.020 0.214** 0.022 0.312** 0.021 0.306** 0.023 Potential slack 0.000 0.001 0.000 0.001 -0.008† 0.005 -0.007 0.005 Unabsorbed slack 0.026** 0.001 0.044** 0.008 0.027** 0.004 0.026** 0.005 Performance below historical
aspirations 2.323** 0.292 0.485** 0.177
Performance above historical
aspirations 0.094 0.122 0.115 0.120
Performance below social
aspirations 0.367 0.335 -0.049 0.177
Performance above social
aspirations 0.201* 0.085 1.182** 0.117
Number of observations 16,741 15,544 15,550 14,133
Number of acquisitions 1,300 1,216 1,288 1,178
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). †. Correlation is significant at the 0.10 level (2-tailed).
This research set out to examine how R&D expenditure could moderate the relationship between performance below aspirations and organizational risk-taking behavior by researching acquisition behavior in US manufacturing firms using panel data collected from a period of 35 years. I proxied both R&D expenditures and acquisitions to be examples of risk-taking behavior, hypothesizing that a make-or-buy decision causes decision-makers to focus on an internal R&D (make) or external acquisition strategy (buy). Earlier performance feedback studies have found that firms may react to performance feedback differently based on organizational characteristics, and therefore I also examined whether a shift-of-focus occurs for low and high R&D-intensity firms. Overall, the results of the data are mostly in line with the predictions. In this section, I first describe the major findings of the study and the theoretical and managerial contributions it entails. Afterwards, I describe the limitations of the study, and pose several suggestions for future research.
Major findings
Innovations can potentially transform firms and industries, but they are also laden with risk (Greve, 2003). Mergers and acquisitions can constitute substantial performance turnaround for firms, but are a complex undertaking and fail often (Zollo & Singh, 2004). Despite the high risk associated with these two highly consequential activities, organizations nonetheless continue to consider both to be viable strategic options.
I first hypothesized (1a) the relationship between performance below aspiration on acquisition count is negative, such that a further decrease relative to aspirations leads to an
increase in acquisition count, but rather found the opposite, namely that a further decrease in performance leads to lower acquisition count. This contradicts the problemistic search argument that implies that an organization would increase search efforts to cater to the performance shortfall, but is in line with findings from Iyer and Miller (2008), who also found such a
contradicting relationship within M&A contexts. I theorize that this contradicting result may be caused by the moderating effect of R&D-intensity, such that decision-makers may increase search efforts to cater to the performance short-fall, but possibly channel those efforts towards other types of risk-taking behavior.
Blonigen and Taylor (2000) argued that R&D-intensity and M&A-activity have an inverse relationship, implying that firms choose between an internal growth strategy with a focus on R&D-expenditure versus an external growth strategy through acquisitions. In numerous performance feedback studies, scholars have considered both R&D-intensity and M&A behavior as outcome variables for risk-taking behavior due to their high consequentiality, but to my
knowledge, no study has combined both concepts in which the one moderates the other. Due to the one being another type of risk-taking behavior it may act as a ‘surrogate’ for the other. I posit that decision-makers are presented with a make-or-buy decision and, following problemistic search, are inclined to consider internal development first as a proxy for local search versus the more ‘distant’ option of external acquisitions. Therefore, this study focused on researching the negative moderating effect of R&D-intensity (internal risk-taking behavior) on the relationship between performance feedback and acquisition likelihood (external risk-taking behavior). My results indeed provide proof for such a negative moderating relationship, for it was present for firms that perform below their historical and social aspiration level. While R&D-intensity and an
acquisition strategy are not necessarily exclusive to each other, and sometimes integrated together, my findings implicate that, rather than complementing, R&D expenditure is seen as a managerial choice of action that is an alternative for acquiring when performing below
aspirations.
For firms that perform above their aspiration level, the moderating effect was only found to be present when attending to the social aspiration level, not on the historical aspiration level. This implies that when firms perform above historical expectations, decision-makers are not urged to change their type of risk-taking behavior, as they are strengthened in their beliefs that the current strategic focus is successful. This is in line with the problemistic search argument: if there is no ‘problem’, managers are less inclined to search and strive for organizational change. Only when considering the performance of fellow competing firms, organizations that perform above the (social) aspiration level have a higher R&D-intensity that acts as a surrogate for acquisition behavior.
Hypothesis 3 theorized that there is a difference in how low and high R&D-intensity firms react on performance feedback , depending on which aspiration type is attended to. In line with my expectation, underperforming firms with low R&D intensity are more likely to attend to historical aspirations than social aspirations when considering an acquisition.
I further hypothesized that underperforming firms with high R&D-intensity would be more inclined to cater to social aspiration than to historical aspirations when considering acquisitions. My results don’t offer proof for this statement, rather indicating that historical aspirations are the only predictor for acquisition behavior for underperforming firms. I posit that historical aspirations are deemed more important by decision makers when considering an
acquisition in underperforming firms, due to managers being better able to analyze historical performance shortfalls and identify its causes, rather than catering to a more ambiguous social performance measure.
For firms that perform above aspirations, only social aspirations were found to be a predictor of acquisition behavior, but stronger for high R&D-intensity firms than for low R&D intensity firms. I posit that this difference may be caused by the nature of high R&D-intensity firms themselves. High R&D-intensity firms already have a natural focus on innovations, as the goal of R&D expenditure is gaining competitive advantage over their competing rivals. This implies that these firms have a tendency to benchmark themselves more explicitly with their rivals, therefore catering to social aspirations more often when considering an acquisition.
In conclude, this study contributes to the existing performance feedback literature in several ways. Most importantly, it supplements the behavioral theory of the firm in acquisition contexts by extending previous work on the subject through analyzing the moderating effect of R&D-intensity on the relationship between performance feedback and acquisition likelihood. It explores the notion that alternative kinds of risk-taking behavior interact with each other and are interdependent. Therefore, it contributes to a finer understanding of the influences of
organizational characteristics on the relationship between performance feedback and organizational risk-taking behavior.
