Historical versus social performance feedback and risk-taking in corporate entrepreneurship : research & developement versus acquisitions

(1)

HISTORICAL VERSUS SOCIAL PERFORMANCE FEEDBACK

AND RISK–TAKING IN CORPORATE ENTREPRENEURSHIP:

RESEARCH & DEVELOPMENT VERSUS ACQUISITIONS

Final Master Thesis Priya Bhageloe Panday Student number: 10362681 MSc Business Administration – Strategy Amsterdam Business School, University of Amsterdam Supervisor: B. (Bernardo) Silveira Barbosa Correia Lima Word count: 14,329 Date: 21/06/2018

(2)

1

Abstract

Prior studies yield contradicting predictions and empirical support of the relationship between performance below aspiration levels and organizational risk–taking. The conflicting views entail on the one hand induced risk–taking when firm decision makers view the

performance gap as repairable and conversely reduced risk–taking when firm decision makers view the performance gap as a threat to firm survival. In response, this study proposes an explanation for the paradoxical views by taking into account different risk–levels of risky search behavior in Corporate Entrepreneurship taken as proxies for risk–taking. I argue that due to the perceived downside risk potential, research & development investments have a relatively low risk–level and acquisition activities have a relatively high risk–level. The findings support the prediction that when performance falls below aspiration levels, firm decision makers become risk–seeking when research & development investments are the proxies for risk–taking. Conversely, no support is found for the expectation of risk–averse behavior in firm decision makers when acquisition activities are the proxies for risk–taking. The findings suggest that differences between the risk–levels of proxies taken to measure risk–taking do affect the results found for the relationship between negative performance feedback and organizational risk–taking behavior. Additionally, I argue that a difference in the benchmarks of social versus historical aspiration levels triggers different mental models in firm decision makers and thus influence the relationship between negative performance feedback and organizational risk–taking behavior. However, the results did not provide support for these final predictions, which entails that the risk–preference of firm decision makers relies on both historical and social aspiration levels.

Key words: performance feedback; historical and social aspiration levels; organizational risk– taking; corporate entrepreneurship; research & development; acquisition

(6)

5

Introduction

Scholars conducted extensive research on how firms learn from past experiences to increase their performance (Argote, McEvily, & Reagans, 2003; Argote & Epple, 1990; Cyert & March, 1963; Garvin, 1993; Haleblian & Finkelstein, 1999; Huber, 1991; Levinthal & March, 1993; Levitt & March, 1988; March, 1991). An extensive stream of research focusses on organizational learning from performance feedback, such as the behavioral theory of the firm (Cyert & March, 1963), prospect theory (Kahneman & Tversky, 1979a) and strategic reference point theory (Fiegenbaum, Hart, & Schendel, 1996). A commonality of these theories is that firms set organizational goals, and performance relative to aspiration levels guides organizational strategic behavior. Primarily, performance exceeding aspiration level results in reduced search behavior and increased strategic perseverance. Conversely,

performance below an organization’s aspiration level signals the existence of a problem that requires attention. Consequently, firms engage in problemistic search behavior to mend performance shortfalls (Cyert & March, 1963; Greve, 2003a; Greve, 2003b).

Numerous studies argue that performance feedback influences organizational risk– taking (e.g. (Audia & Greve, 2006; Bromiley, 1991; Chen, 2008; Greve, 1998; Kacperczyk, Beckman, & Moliterno, 2015; March & Shapira, 1992). However, these studies yield

inconsistent findings regarding the effects of negative performance feedback on risk–taking. Various studies argue in line with the behavioral theory of the firm and prospect theory that when firms perform below their aspiration levels, they become more risk–seeking, due to increased willingness of enacting change to overcome their performance shortfalls (Audia & Greve, 2006; Boyle & Shapira, 2012; Bromiley, 1991; Cyert & March, 1963; Greve, 1998; Greve, 2011; Miller & Chen, 2004; Palmer & Wiseman, 1999; Wiseman & Bromiley, 1996). On the contrary, other studies argue in line with the threat of rigidity theory that when firms perform below their aspiration levels, they become more risk–averse by becoming reluctant to

(7)

6 change and restricting activities to their core business (Audia & Greve, 2006; McNamara & Bromiley, 1997; Miller & Bromiley, 1990; Sitkin & Pablo, 1992; Staw, Sandelands, & Dutton, 1981). These conflicting findings could be due to differences in boundary conditions in examining the relationship between negative performance feedback and organizational risk–taking preferences (Audia & Greve, 2006; Desai, 2008; Kim, Finkelstein, & Haleblian, 2015; Lim, 2015). For example, Audia and Greve (2006) argue that small firms tend to be more risk–averse when performance drops below aspiration level, whereas this effect or increased risk–taking is not in place for large firms. Another example is Desai (2008), who argues limited operating experience and poor legitimacy lead to risk–aversion in organizations following negative performance feedback. However, these studies do not take unobserved heterogeneity in the outcomes into account, as prior research uses various organizational outcomes as proxies for risk–taking.

Differences in risk–levels of the outcomes used to measure risk–taking, could affect the relationship between negative performance feedback and organizational risk–taking, due to the relatively low– versus high downside potential associated with different organizational activities (March & Shapira, 1987). Hence, to obtain a better understanding of the relationship between negative performance feedback and organizational risk–taking propensity, I include the heterogeneity of risk–levels in different types of problemistic search activities. I consider insights from Corporate Entrepreneurship literature on differences in relative risk–levels between inherently risky entrepreneurial activities.

Corporate entrepreneurial behavior entails firms setting up new businesses within their organization via innovation or corporate venturing, herewith strategically renewing the firm through new combinations of resources, to improve their long–term performance (Bierwerth, Schwens, Isidor, & Kabst, 2015; Dess et al., 2003; Guth & Ginsberg, 1990; MacMillan & Day, 1987; Zahra & Covin, 1995). Corporate entrepreneurship (CE) is an inherently risky

What are do these boundary conditions capture? The issue is heterogeitey not boundary conditions per se

Why?

How is this related to the issue above?

(8)

7 undertaking of firms (Burgelman, 1983a; Guth & Ginsberg, 1990; Sharma & Chrisman, 2007; Zahra & Covin, 1995), because it requires resources and increases the variance of possible outcomes (March & Shapira, 1987) to ultimately increase their organizational outcomes such as value and firm performance (Burgelman, 1983b; Zahra, 1995; Zahra, 1996). CE generates knowledge, which allows firms to learn, adapt and respond to challenges in their environment (Zahra, Nielsen, & Bogner, 1999). Consequently, this increases a firm’s competitive

advantage (Dess et al., 2003; Ireland, Covin, & Kuratko, 2009). Hence, CE is a form of organizational search for new knowledge and resources which increase firm performance.

Scholars have distinguished between many types and approaches of CE behaviors, that are all oriented towards rejuvenating, redefining or establishing innovation and can appear simultaneously in different business units of one firm (Covin & Miles, 1999; Dess et al., 2003; Wolcott & Lippitz, 2007). Depending on whether the conception of the new idea and development of required resources occurs internal– or external of the firm, CE distinguishes internal– from external corporate venturing (Basu & Wadhwa, 2013). Internal corporate venturing entails experimental organizational learning and refers to the conception of the new business idea and development of required capabilities within the firm, e.g. via research & development (Basu & Wadhwa, 2013; Covin & Miles, 2007; Miles & Covin, 2002; Schildt, Maula, & Keil, 2005). Conversely, external corporate venturing entails acquisitive and vicarious learning and refers to an inter–organizational relationship with the purpose to either create new ventures or develop existing internal business via e.g., corporate venture capital investments, alliances, joint ventures or acquisitions.

This study attempts to reconcile the paradoxical views regarding the relationship between negative performance feedback and organizational risk–taking behavior, by taking into account the different risk–levels of outcomes in corporate entrepreneurial behavior. Due to the relative difference in risk–levels of internal versus external corporate entrepreneurship,

(9)

8 I take research & development investments relative to acquisition activity as proxies for respectively low and high organizational risk–taking. Consequently, when performance falls below aspiration levels, I expect firms to engage more in research & development because of the relatively low potential downside risk associated with this type of search. This prediction is in line with the behavioral theory of the firm and prospect theory, which posit that firm decision makers increase risk–taking to mend performance shortfalls. On the other hand, when performance falls below aspiration levels, I expect firms to engage less in acquisition activities because of the relatively high potential downside risk associated with this type of search. This prediction is in line with the threat of rigidity theory, which posits that firm decision makers to be more risk–averse when performance falls below aspiration levels.

Additionally, scholars have recently started to consider different effects of the type of aspiration levels. Aspiration levels are based on two distinct sources of performance feedback: historical and social. Historical performance feedback relies on the focal firm’s past

performance, whereas social performance feedback relies on the performance of similar firms in a reference group (Cyert & March, 1963; Greve, 2003b). Generally, scholars studying the effects of performance feedback aggregate historical and social performance feedback and treat the two as eliciting similar strategic behavior (Audia & Greve, 2006; Cyert & March, 1963; Greve, 2011; Wiseman & Bromiley, 1996). In the second part of this study, a different perspective is taken in which I consider historical and social performance feedback to

encourage different strategic behavior. This is due to differences in reference points which provide a firm’s decision makers with different informative content and are thus interpreted differently (Greve, 2003b; March & Shapira, 1987). Historical aspiration levels rely on accessible information created within the organization, contrast an organization’s current performance with its past performance and indicate how well an organization could perform

(10)

9 (Greve, 2003a, p 42). Social aspiration levels rely on ambiguous external performance

information, contrast an organization’s current performance to the performance of similar others and indicate how well an organization should perform (Cyert & March, 1963). Despite these fundamental differences in the respective underlying benchmarks, only limited research is done on their different effects on strategic firm behavior.

In seeking to advance performance feedback theory, this study considers insights from corporate entrepreneurship to understand better how different reference points in performance feedback below aspiration levels affect strategic organizational behavior. I posit that different reference points in learning from performance feedback activate different mental frameworks in the decision makers of the firm, due to their inward– (historical) and outward (social) looking nature, which affect firm engagement in either internal– or external corporate venturing.

Performance below the historical aspiration level signals an internal performance shortcoming relative to its past performance (Greve, 2003b). This activates an inward–looking model in firm managers, who will engage in local problemistic search for a solution.

Consequently, when performance falls below the historical aspiration level, firm decision makers are likely to engage in internal search for a solution, via internal corporate venturing by increasing research & development activities (Burgelman, 1983b; Wolcott & Lippitz, 2007). Therefore, when performance falls below aspiration levels, I expect the increase in research & development (R&D) investments to be stronger for performance below historical aspiration level than performance below social aspiration level.

On the other hand, performance below the social aspiration level signals external performance shortcomings relative to similar peers in an industry (Greve, 2003b). This activates an outward–looking model in firm managers, who will engage in external

problemistic search for a solution. Consequently, when performance falls below the social Text

(11)

10 aspiration level, firm decision makers are likely to engage in external search for a solution, via external corporate venturing by increasing corporate venture capital investments, alliances, joint ventures or acquisitions (Schildt et al., 2005). Therefore, when performance falls below the social aspiration level, I expect the decrease in acquisition activity to be less than when performance falls below the historical aspiration level.

This study considers insights from corporate entrepreneurship to gain a better understanding of the relationship between performance feedback and risk–taking, as well as how different reference points in performance feedback below aspiration levels affect organizational behavior. In general, this paper makes three contributions to the existing literature. Firstly, it adds to previous work on the debate of the relationship between negative performance feedback and organizational risk–taking behavior, by considering heterogeneity in levels of risk of the organizational outcomes. Secondly, it contributes to performance feedback theory by shedding light on the different effects of social and historical reference points on consequent strategic behavior. Thirdly, by examining antecedents of internal– and external corporate venturing separately, this study advances the literature on corporate entrepreneurship.

Theory and Hypotheses – I

The following section provides a review of the central insights in existing literature on performance feedback, organizational risk–taking behavior and corporate entrepreneurship (CE), and presents the first two hypotheses of this study. Firstly, the nature of performance feedback and its influence on firm strategic behavior, such as search and risk–taking is

introduced. Then, I introduce corporate entrepreneurship, build on its insights to fill in the gap and present the first research question. Subsequently, the effects of negative performance

(12)

11 feedback on internal CE are outlined, leading to the first hypothesis. Finally, the effects of negative performance feedback on external CE are outlined, leading to the second hypothesis.

Performance Feedback and Organizational Risk–Taking Performance Feedback and Organizational Search Behavior

Firms learn from past experiences to increase their performance (Argote et al., 2003; Argote & Epple, 1990; Cyert & March, 1963; Garvin, 1993; Haleblian & Finkelstein, 1999; Huber, 1991; Levinthal & March, 1993; Levitt & March, 1988; March, 1991). In particular, systematic performance evaluation enables organizational learning through reconceptualizing and gradually changing routines based on past performance feedback (Feldman & Pentland, 2003; Levitt & March, 1988). Organizational learning from performance feedback improves firms decision making quality, which in turn leads to increased performance (Huber, 1991; Moynihan & Landuyt, 2009). The focus of an extensive stream of research on performance feedback is the influence of performance relative to aspiration level on firm strategic behavior (Audia & Greve, 2006; Bromiley, 1991; Greve, 2003a; Greve, 2003b). The foundations of this research are in the behavioral theory of the firm (Cyert & March, 1963). Cyert and March (1963) pose that firms learn from past experiences. This entails that decision makers within the firm set organizational goals, id est aspiration levels, which serve as a benchmark to evaluate past performance.

However, according to the behavioral theory of the firm, managers cannot process all available information in performance evaluation because they are cognitively restrained by bounded rationality (Cyert & March, 1963; March & Simon, 1958). Therefore, managers simplify evaluation by reflecting on the achievement of the organization’s performance objectives when making strategic decisions (e.g.(Bromiley, 1991; Cyert & March, 1963;

(13)

12 Fiegenbaum, 1990; Greve, 1998; Greve, 2003a; Greve, 2003b; Jordan & Audia, 2012;

Kahneman & Tversky, 2013; Nielsen, 2014; Parker, Krause, & Covin, 2017; Shinkle, 2012). The discrepancy between aspiration level and firm performance guides strategic behavior, by altering managerial risk preferences and organizational search behavior (Cyert & March, 1963; Fiegenbaum, 1991). Managers use information on positive– versus negative attainment discrepancy of goals determined by the organization to control search and decision making (Cyert & March, 1963; Iyer & Miller, 2008; Lim, 2015; Miller & Chen, 2004).

Performance feedback above aspiration levels is categorized as a gain and allows the firm to exploit its effective strategies (March & Simon, 1958; March, 1991), whereas performance below

aspiration level is categorized as a loss and calls for problemistic search behavior towards

potentially risky organizational change (Cyert & March, 1963; Greve, 2003a; Greve, 2003b; March & Simon, 1958). Cyert and March define problemistic search as “search that is stimulated by a problem … and is directed toward finding a solution to that problem” (1963, p 121). Furthermore, the problemistic search continues until a solution for performance shortfalls is found and can be initiated again when the problem reoccurs (Greve, 2003b).

Performance Below Aspiration Level and Risk–Taking

Prior research argues that performance below aspiration levels affects organizational risk preference (e.g. (Audia & Greve, 2006; Bromiley, 1991; Chen, 2008; Greve, 1998; Kacperczyk et al., 2015; March & Shapira, 1992; Miller & Chen, 2004). One widely adopted definition of risk is the ‘variation in the distribution of possible outcomes, their likelihoods, and their subjective values.’ (March & Shapira, 1987, p. 1404). Risk increases when the variance of the probability of possible positive or negative outcomes increases, which entails that the outcome becomes less predictable. However, managers are argued to be more

(14)

13 organizational outcome seems more salient to them (March & Shapira, 1987; Miller &

Leiblein, 1996). Bromiley, Miller and Rau thus define risk as the “downside unpredictability of business outcome variables” (Bromiley, Miller, & Rau, 2001); p. 261). This study adopts the latter definition of risk as the probability of downside outcomes. Several theories build on the performance feedback theory to explain consequent organizational risk–taking behavior. Scholars have argued that when performance exceeds aspiration levels, firm managers become more risk–averse, due to the absence of the need to alter their practices and increase strategic perseverance (Greve, 1998; Greve, 2003a; March & Shapira, 1987). However, contradictory findings indicate an ambiguous relationship between performance feedback and risk–taking behavior when performance falls below aspiration levels.

Various studies argue that when firms perform below their aspiration levels, they become more risk–seeking, due to increased willingness of enacting change to overcome their performance shortfalls (Audia & Greve, 2006; Boyle & Shapira, 2012; Bromiley, 1991; Cyert & March, 1963; Greve, 1998; Greve, 2011; Miller & Chen, 2004; Palmer & Wiseman, 1999; Wiseman & Bromiley, 1996). These findings are in line with the behavioral theory of the firm (Cyert & March, 1963), prospect theory (Kahneman & Tversky, 1979a; Kahneman &

Tversky, 1979b) and strategic reference point theory (Fiegenbaum & Thomas, 1995). According to prospect theory, individuals can evaluate the same opportunity differently, depending on the perceived potential gain or loss (Kahneman and Tversky, 1979; 1986). The theory suggests that decision making among risky alternatives is based on potential values of gains and losses relative to the status quo. Consequently, managerial behavior is risk–averse when performance is above their aspiration level (probability of loss is higher), and risk– seeking when performance is below their aspiration level (probability of gain is higher) (Kahneman & Tversky, 1979a; Tversky & Kahneman, 1992; Wehrung, 1989). Risky alternatives become increasingly acceptable to firm decision makers when performance is

(15)

14 below aspiration level because the firm enters the loss domain(Greve, 2003b). Building on prospect theory is strategic reference point theory, which states that firms are driven by benchmarking, and top managers set strategic reference points (SRP) as organizational goals for their organization. A firm’s strategic choice behavior depends on performance relative to its SRPs, being risk–averse when past performance is above its SRP, and to an even greater extent risk–seeking when past performance is below its SRP (Fiegenbaum, 1990;

Fiegenbaum, 1991).

On the contrary, other studies argue that when firms perform below their aspiration levels, they become more risk–averse by becoming reluctant to change and restricting

activities to their core business (Audia & Greve, 2006; McNamara & Bromiley, 1997; Miller & Bromiley, 1990; Sitkin & Pablo, 1992). These findings are in line with the threat of rigidity theory (Staw et al., 1981). In contrast to the behavioral theory of the firm and prospect theory, the threat of rigidity theory posits low performance is perceived as a threat to organizational survival and not as a repairable gap (Ocasio, 1995; Sitkin & Pablo, 1992; Staw et al., 1981). Negative performance feedback threatens the firm and leads to anxiety, psychological stress, and restriction on information processing in managers. Under these conditions, managers are restricted in generating and considering risky solutions, which makes them risk–averse (Staw et al., 1981). Additionally, managers experience a conflict between the desire to improve performance by making risky decisions and avoid additional losses by taking fewer risks (Lopes, 1987). The managerial cognitive constraints on generation and evaluation of risky solutions, combined with the desire to avoid additional losses imply that performance below aspiration levels enhances risk–aversion in firm decision makers.

In sum, the conflicting theoretical predictions in prior research of negative

(16)

15 for both camps. On the contrary, prior research consistently substantiates that positive

performance feedback enhances risk–averse behavior.

Performance Feedback and Risk–Taking:

Literature Gap, Corporate Entrepreneurship and Research Question Literature Gap

Several theories have often been applied to explain risk–taking behavior of firms following performance feedback. Both conflicting theoretical predictions are supported by different empirical evidence, which calls for an examination of boundary conditions to the relationship between negative performance feedback and organizational risk–taking behavior. March and Shapira (1987, 1992) proposed a shifting–focus model of organizational risk– taking, which attempts to reconcile the opposing views on organizational risk–taking behavior following negative performance feedback. They posit that managers do not rely on a single reference point, but switch their focus to reference points between the aspiration level versus organizational survival. In turn, this influences the interpretation of performance feedback and elicits different strategic behavior in firm decision makers regarding increasing or decreasing risk–taking. In the situation of performance below aspiration level, but above the

organizational survival point, there are two possible scenarios. In the first scenario managers focus on the aspiration level. This implies that they view the performance gap as repairable and, in their attempt to repair the gap, become more risk–seeking as performance falls below aspiration level. In the second scenario managers focus on the organization’s survival point. This implies that they interpret falling performance as a threat to firm survival. Consequently, managers become risk–averse, because they are cognitively restricted in evaluating possible risky alternatives or wish to avoid bankruptcy caused by additional losses (Audia & Greve, 2006; March & Shapira, 1987; March & Shapira, 1992).

(17)

16 Several scholars have tried to settle the debate by building on the shifting focus theory, and considering various properties and firm conditions as moderators of the relationship (Audia & Greve, 2006; Desai, 2008; Kim et al., 2015; Lim, 2015). For example, firm size can be a predictor for contrary findings, because managers in small organizations with low

resources are more likely to perceive negative performance feedback as a threat to firm survival, instead of as a repairable gap (Audia & Greve, 2006; Greve, 2011). Furthermore, firms become risk–averse when they are proximate to bankruptcy, which is in line with the threat of rigidity theory. Another example comes from Desai (2008), who argues limited operating experience and poor legitimacy lead to risk–aversion in organizations following negative performance feedback. However, prior studies did not take into account unobserved heterogeneity in the various outcomes used as proxies to measure organizational risk–taking. Differences in risk–levels of the outcomes used to measure risk–taking, could affect the relationship between negative performance feedback and organizational risk–taking, due to the relatively low– versus high downside potential associated with different organizational activities (March & Shapira, 1987). Hence, to obtain a better understanding of the relationship between negative performance feedback and organizational risk–taking propensity, in this study I include the heterogeneity of risk–levels in different types of problemistic search activities. In seeking to fill the gap, I consider insights from Corporate Entrepreneurship literature on differences in relative risk–levels between inherently risky entrepreneurial activities.

Corporate Entrepreneurship

Over the past decades, entrepreneurial activities within existing organizations have been increasingly studied (Birkinshaw, 1997; Burgelman, 1983a; Burgelman, 1983b; Covin & Miles, 1999; Sharma & Chrisman, 2007; Wolcott & Lippitz, 2007; Zahra & Covin, 1995;

(18)

17 Zahra, 1995; Zahra, 1996; Zahra et al., 1999; Zahra, 2015). Corporate entrepreneurship (CE) is an increasingly studied and numerously defined concept, one of which is that CE is “…the process whereby an individual or a group of individuals, in association with an existing organization, create a new organization or instigate renewal or innovation within that organization” (Sharma & Chrisman, 2007), p. 18). Companies engage in CE to create new knowledge, which enables acquisitive and experimental learning (Keil, 2004; Kuratko, Ireland, & Hornsby, 2001). This form of organizational search allows firms to renew themselves, adapt to challenges in their environment, fuel growth, enhance profitability and increase their competitive advantage (Covin & Miles, 1999; Keil, 2004; Zahra & Covin, 1995; Zahra et al., 1999). CE consists of internal– and external corporate venturing,

depending on whether the conception of the new idea and development of required resources occurs internal– or external of the firm (Basu & Wadhwa, 2013). Internal corporate venturing entails experimental organizational learning and refers to the conception of the new business idea and development of required capabilities within the firm (Basu & Wadhwa, 2013; Covin & Miles, 2007; Schildt et al., 2005). This can be transforming research & development activities into innovation and new business incubation (Burgelman, 1983b; Narayanan, Yang, & Zahra, 2009). Conversely, external corporate venturing entails acquisitive and vicarious learning, and refers to an inter–organizational relationship with the purpose to either create new ventures or develop existing internal business, e.g. via corporate venture capital (CVC) investments, alliances, joint ventures or acquisitions (Schildt et al., 2005). Companies are engaging in external corporate venturing, to leverage a partner in an equity or non–equity inter–organizational relationship to learn from external knowledge sources (Keil, 2004; Miles & Covin, 2002; Schildt et al., 2005). CVC investments are minority equity investments made by established companies in privately held ventures (Basu & Wadhwa, 2013). Strategic alliances and joint ventures are agreements between two or more companies to pool physical

(19)

18 and human resources, access new markets and supply sources, capitalize on technology, deploy assets better, and become more profitable (Walters, Peters, & Dess, 1994). Corporate acquisitions involve a company buying another company for similar reasons, yet the other party need not consent to this takeover. Several studies argue that managers adopt new routines and strategies such as innovations and acquisitions to overcome performance shortfalls and increase organizational growth (Audia & Greve, 2006; Greve, 2003a; Greve, 2003b; Greve, 2011; Haleblian & Finkelstein, 1999; Iyer & Miller, 2008). Prior research has shown that ICV in the form of research & development can serve to some extent as a

substitute for ECV, such as corporate acquisitions or joint ventures (Friedman, Berg, & Duncan, 1979; Hall, 1988).

CE is an inherently risky undertaking of firms (Burgelman, 1983a; Guth & Ginsberg, 1990; Sharma & Chrisman, 2007; Zahra & Covin, 1995), because it requires resources and increases the variance of possible outcomes (March & Shapira, 1987), to ultimately increase their organizational outcomes such as value and firm performance (Burgelman, 1983b; Zahra, 1995; Zahra, 1996). However, not all CE activities have the same risk–levels because some entrepreneurial activities have a high probability of failure (Zahra & Covin, 1995). When comparing research & development investments to acquisition activities, there are

fundamental differences in the associated downside risk. Research & development

investments are relatively low, flexible and firm–specific, which implies less probability of having a detrimental impact on the organization. Acquisition activities require high

investments, which are rigid and often fail to create value (King, Slotegraaf, & Kesner, 2008). Thus, acquisition activity as a response to low performance are more likely to have a

detrimental impact on the organization, as it is likely to move the firm closer to bankruptcy (March & Shapira, 1992). In this study, I consider these fundamental differences and posit that in CE behavior, research & development investments have substantially less downside

(20)

19 risk than acquisition activities. Building on the shifting focus model by March and Shapira (1992), I posit that when performance falls below aspiration levels, firms are risk–seeking when the outcome used to measure risk is research & development investments due to the relatively low downside risk. Adversely, when performance falls below aspiration levels, I expect that firms are risk–averse when the outcome used to measure risk is acquisition activities due to the relatively high downside risk.

Research Question I

This study aims to provide an in–depth understanding of the mechanisms of performance below aspirations on organizational risk–taking behavior, in light of ICE and ECE. Therefore, the research question this study aims to answer is as follows:

What is the influence of R&D investments versus Acquisition Activities as proxies of Risk–Taking, on the relationship between performance below aspiration level and Organizational Risk–Taking?

Performance Feedback and Research and Development investments

Internal corporate venturing refers to the process in which a firm “transforms research & development activities at the frontier of corporate technology into new businesses”

(Burgelman 1983b, p 223). Research & development intensity (R&D expenses divided by sales) reflects the extent to which a firm chooses to internally develop new processes or products (Miller & Bromiley, 1990), and is an inherently risky undertaking of firms due to the uncertainty of the innovation outcomes (Baird, 1986). In the light of CE activities, research & development investments are relatively low stake, flexible in initiation and retraction, and firm–specific, thus do not rely on stakeholders external to the focal firm. Furthermore, research & development investments have greater secrecy on strategic novel internal developments and therefore lower the imitability of competitive capabilities by competitors (Peteraf, 1993). This implies that of the CE activities, R&D investments have a low

(21)

20 probability of having a detrimental impact on the organization (March & Shapira, 1992). This is in line with the behavioral theory of the firm and prospect theory. When performance falls below aspiration levels, I expect firms to engage more in problemistic search via research & development investments, due to the relatively low downside risk–level associated with research & development activities. Following this reasoning, hypothesis 1 is as follows:

H1: When performance is below the aspiration level, performance decreases lead to more risk– taking concerning R&D investments.

Performance Feedback and Acquisition Activity

External corporate venturing refers to corporate venture capital (CVC) investments, alliances, joint ventures and acquisitions of entrepreneurial ventures, and is used by

companies to integrate knowledge sources and technologies from beyond the boundaries of their firm (Covin & Miles, 2007; Miles & Covin, 2002; Schildt et al., 2005). Several studies argue that managers adopt new routines and strategies such as innovations and acquisitions to overcome performance shortfalls and increase organizational growth (Audia & Greve, 2006; Greve, 2003a; Greve, 2003b; Greve, 2011; Haleblian & Finkelstein, 1999; Iyer & Miller, 2008). Acquisitions reflect the focal firm’s managerial intention because the target does not need to consent to merge in e.g. a hostile takeover. Acquisition activities are inherently risky, because post–merger performance improvement is uncertain and they entail high integration, high rigidity and high investment (Schildt et al., 2005). In the light of CE activities,

acquisitions are relatively high stake, rigid, and often fail to create value due to, among other reasons, the cultural and operational differences (King et al., 2008). Knowledge of acquiring a target is available to competitors of the firm, which discloses to some extent the strategic plan of the focal firm. Therefore, engaging in an acquisition increases imitability by competitors (Peteraf, 1993). This implies that, of the CE activities, acquisitions have a relatively high

(22)

21 probability of having a detrimental impact on the organization, as it moves the firm closer to bankruptcy (March & Shapira, 1992). This line of reasoning is compatible with the threat of rigidity theory. When performance falls below aspiration levels, I expect firms to engage less in problemistic search via acquisitions, due to the relatively high downside risk–level

associated with acquisition activities. Following this reasoning, hypothesis 2 is as follows: H2: When performance is below the aspiration level, performance decreases lead to fewer risk– taking concerning Acquisition Activities.

Theory and Hypotheses – II

The following section provides a review of the central insights in existing literature on the role of different reference points in performance below aspiration level on strategic organizational behavior, and builds on insights in corporate entrepreneurship (CE) to present the last two hypotheses of this study. On the one hand, the effects of social– versus historical aspiration levels on internal CE are outlined, leading to the third hypothesis. On the other hand, the effects of social– versus historical aspiration levels on external CE are outlined, leading to the fourth hypothesis. Figure 1 displays the conceptual model of this study.

Historical– versus Social Performance Feedback – Literature Gap and Research Question This section discusses the role of different reference points in the relationship between performance below aspiration level and research & development investments, as well as acquisition activity.

As discussed in the section of performance feedback and organizational search, performance below an organization’s aspiration level signals the existence of a problem that requires attention. Consequently, firms engage in problemistic search behavior to mend performance shortfalls (Cyert & March, 1963; Greve, 2003a; Greve, 2003b). Conversely,

(23)

22 performance exceeding aspiration level results in reduced search behavior and increased strategic perseverance.

Literature Gap

Performance feedback relative to aspiration levels differs in reference points, id est historical and social (Cyert & March, 1963). Generally, scholars studying the effects of performance feedback aggregate historical and social performance feedback and treat the two as eliciting similar strategic behavior (Audia & Greve, 2006; Cyert & March, 1963; Greve, 1998; Greve, 2003a; Wiseman & Bromiley, 1996). However, historical and social reference points differ in their respective underlying benchmarks and thus allow for differences in organizational learning, as they generate different information on performance and are evaluated through dissimilar cognitive and organizational processes (Greve, 2003b; Herriott, Levinthal, & March, 1985; Kim et al., 2015). Historical performance feedback relies on the focal firm’s past performance, whereas social performance feedback relies on the

performance of similar firms in a reference group (Cyert & March, 1963; Greve, 2003b). Historical aspiration levels incorporate organizational factors of past performance that reflect managerial capabilities and resources, thus indicate how well the organization could perform and incentivizes experimental learning from past experience (Greve, 1998; Greve, 2003b; Kim et al., 2015). Social aspiration levels incorporate environmental factors of relative performance of a reference group and allow for benchmarking, thus indicate how well the organization should perform, because stakeholders expect the firm to perform at least as well as similar others, and incentivize acquisitive and vicarious learning (Fiegenbaum & Thomas, 1995; Greve, 1998; Greve, 2003b; Kim et al., 2015).

In sum, historical and social aspiration levels differ in reference points which provide a firm’s decision makers with different informative content and are thus interpreted

(24)

23 relationship between negative performance feedback and strategic organizational behavior in different ways, depending on the combination of the nature of the performance feedback and the organizational outcome. Despite these fundamental differences, historical and social aspirations are frequently combined to form one aspiration level (Audia & Greve, 2006; Cyert & March, 1963; Greve, 1998; Greve, 2003a; Greve, 2011), and only limited research is done on their different effects on strategic firm behavior.

This study takes on a novel perspective, in which I consider historical and social performance feedback to encourage different strategic behavior. In seeking to advance performance feedback theory, this study considers insights from internal and external

corporate entrepreneurship to understand better how different reference points in performance feedback below aspiration levels affect strategic organizational behavior. I posit that different reference points in learning from performance feedback activate different mental frameworks in the decision makers of the firm, due to their inward– (historical) and outward (social) looking nature. When performance falls below aspiration levels, I expect the increase in research & development investments (internal/local search) to be stronger for performance below historical aspiration level than performance below social aspiration level. Furthermore, when performance falls below aspiration levels, I expect the decrease in acquisition activities (external) to be less for performance below social aspiration level than performance below historical aspiration level.

Research Question II

This study aims to advance performance feedback theory by considering the effect of

different reference points in aspiration levels on research & development investments (local search) and acquisition activities (external search). Therefore, the second research question this study aims to answer is as follows:

(25)

24 What is the role of aspiration type between performance below aspiration level and

Research and Development investments and Acquisition Activities?

Historical– versus Social Performance Feedback and Research and Development investments Differences in how firm decision makers interpret historical– versus social aspiration attainment discrepancy, influences their subsequent strategic behavior (Kim et al., 2015). Performance below historical aspiration levels activates internally focused learning models in firm decision makers to overcome internal performance shortfalls relative to past performance (Greve, 2003b). This activates an inward–looking model in firm managers, who are more likely to engage in internal problemistic search for a solution. Conversely, performance below social aspiration levels activates acquisitive and vicarious learning models in firm decision makers to overcome external performance shortfalls relative to similar others (Fiegenbaum & Thomas, 1995; Greve, 1998; Greve, 2003b; Kim et al., 2015). This activates an outward– looking model in firm managers, who are more likely to engage in external problemistic search for a solution.As proposed in the first hypothesis, following negative performance feedback I expect firms to engage more in risky problemistic search via research & development investments, due to the relatively low downside risk–level associated with research & development activities. When firms engage in research & development activities to mend performance shortfalls, this reflects internal problemistic search for a solution. Hence, I expect managers to engage more in internal search for a solution through research & development when performance falls below the historical aspiration level, than when

performance falls below the social aspiration level. Following this reasoning, hypothesis 3 is as follows:

(26)

25 H3: The negative relationship between performance below Historical aspiration level and R&D investments is more negative than the relationship between performance below Social aspiration level and R&D investments.

Historical– versus Social Performance Feedback and Acquisition Activity

As proposed in the second hypothesis, following negative performance feedback I expect firms to engage less in risky problemistic search via acquisition activities, due to the relatively high downside risk–level associated with acquisition activities. However, when performance falls below social aspiration levels, this activates an outward–looking model in firm managers, who are then more likely to engage in external problemistic search for a solution. Acquisition activities reflect a type of external search to improve organizational performance. Hence, when performance falls below the social aspiration level, I expect the decrease in acquisition activities to be less than when performance falls below the historical aspiration level. Following this reasoning, hypothesis 4 is as follows:

H4: The positive relationship between performance below Social aspiration level and Acquisition Activities is less positive than the relationship between performance below Historical aspiration level and Acquisition Activities.

(27)

(28)

27

Method

The following section explains the research approach and design of this paper. Firstly, the sampling strategy is discussed. Then, a detailed operationalization of the independent, dependent and control variables is given. Finally, details on the models used to analyze the data are given.

Sample and Data

This study uses combined secondary data of publicly traded U.S. companies, gathered from the WRDS Compustat database for financial– and industry data, and from Thomson One’s SDC Platinum database for information on firms’ acquisition activity. The panel is restricted to high–tech industries because high–tech industries are highly innovative and therefore search behavior is an appropriate response to low performance. Furthermore, previous research has indicated firms in high–tech industries engage more in corporate

entrepreneurship than firms in moderately changing environments (Blonigen & Taylor, 2000), and in particular engage in acquisition activity to increase organizational performance (Ahuja & Katila, 2001; Desyllas & Hughes, 2010). Firms operating in high–technology industries have high–growth potential due to the disruptive nature of technological advancements, as well as high risk due to the inherent uncertainty of company value relying on future outcomes or developments in unproven, uncharted fields (Kohers & Kohers, 2000). The innovation of new technology, product or service entails heterogeneity and gives a company a competitive advantage (Barney, 1991; Peteraf, 1993). However, innovation is subject to imitation and therefore companies in rapidly changing environments, such as high–technology industries, need to renew themselves via innovation through research & development or acquisitions to stay ahead.

(29)

28 The choice for the manufacturing industry is to mitigate difficulties produced by using accounting data from different kinds of businesses and to allow comparison with previous studies (Blonigen & Taylor, 2000; Desyllas & Hughes, 2010; Hayton, 2005). This resulted in a set of high–tech manufacturing firms, with their primary activity in SIC 357 – Computer and Office Equipment, in the period 1980–2014. The independent performance variables and the control variables lag the dependent variables by one year. This entails that the data of performance and control variables ranged from the years 1979 to 2013, whereas the corresponding dependent variables ranged from the years 1980 to 2014.

Independent Variable – Performance Feedback

Return On Assets (ROA = net income divided by total assets) is used as a measurement for firm performance, as it is an overall performance measure, corresponds better with top– level organizational goals and the subsequent decisions of firm decision makers (Gavetti, Greve, Levinthal, & Ocasio, 2012). Furthermore, ROA is a commonly used measure of firm profitability within manufacturing industries and it is applied in substantial prior research, which provides comparability to many other studies on the effects of performance feedback on organizational behavior (Bromiley, 1991; Greve, 1998; Greve, 2003a; Kim et al., 2015). Firm performance is evaluated against two levels: historical– and social. The historical aspiration level relies on the focal firm’s past performance and is calculated as the average ROA of the firm’s past three years. The social aspiration level relies on past performance of similar peers and the average of other firms' performance, calculated as the mean ROA of all similar companies except for the firm of interest (Cyert & March, 1963), defined at the three– digit SIC level.

Firm’s performance feedback is computed by taking actual performance at year t – 1, and aspiration level at year t – 2. Historical performance feedback is calculated by subtracting the

(30)

29 firm’s average ROA of the preceding three years from the focal year's ROA. Social

performance feedback is constructed by subtracting the mean ROA of all companies within the industry from the firm’s ROA in the same year. To create separate variables for

performance below and above (historical and social) aspiration level, spline variables were created where performance below aspiration entails all observations are smaller than zero, and performance above aspiration entails all observations are higher than zero (Desai, 2008).

Dependent Variables – Corporate Entrepreneurship

Research and Development Intensity – Research & development expenditures divided by sales is used as a proxy for search intensity (Chen & Miller, 2007). Research & development intensity reflects firm–specific search expenditures adjusted for firm size. All cases were excluded where research & development intensity is above 1 and supersedes sales, as these are firms that are likely to have different incentives to engage in excessive research & development other than past performance.

Acquisition Counts – Acquisition counts reflect the frequency of acquisitions made by a company in one fiscal year. This variable is a count variable, which implies it consists of non– negative integer values, not containing decimals (Cameron & Trivedi, 2013). The counts reflect acquisition announcement date because this reflects the managerial intention of the focal firm and is not depending on external factors that could prevent the takeover from taking place after the announcement.

Control Variables

Firm size – Risk–taking propensity might increase along with firm size, due to the lower levels of perceived risk in a firm with many resources (Audia & Greve, 2006; Greve, 2011). The measure for firm size is constructed as the logged number of employees because

(31)

30 this specification captures the effect of size on risk–taking better. This entails that a given percentage of increase has the same effect regardless of firm size (Audia & Greve, 2006; Greve, 2011). Furthermore, in the mergers and acquisitions literature, firm size is often found to positively correlate with the probability of acquiring (Tremblay & Tremblay, 1988, in Bloningen & Taylor, 2000).

Distance from Bankruptcy – Firm’s proximity to bankruptcy influences risk–averse behavior, due to the increased likelihood of detrimental results (March & Shapira, 1992). Therefore, I control for closeness to bankruptcy by including the Altman’s Z measure. The distance from bankruptcy with Altman’s Z–score is measured as: (3.3 x income before interest expense and taxes divided by total assets) + (1.4 x retained earnings divided by total assets) + (1.2 x working capital divided by total assets) + (1.0 x sales divided by total assets) + (0.6 x market value of equity divided by total liability). A lower value of the Altman’s Z measure indicates a higher risk of bankruptcy.

Organizational Slack – Organizational slack might influence risk–taking, due to the financial resources it provides and thus perceived buffer for adverse outcomes. Following prior research, this study adopts two proxies for organizational slack (Chen & Miller, 2007; Iyer & Miller, 2008; Lim & McCann, 2013). Absorbed slack is the working capital to sales ratio and unabsorbed slack is measured as the current ratio (current assets divided by current liabilities).

Environmental Dynamism – Environmental Dynamism is the rate of change and the degree of instability of factors within an environment (Simerly & Li, 2000). When

environmental dynamism is high, the effect of negative social performance feedback on corporate entrepreneurial behavior could be more impactful for highly dynamic environments than in moderately dynamic environments because highly dynamic environments lower the relevance of past capabilities and strategies (Greve, 2003b). This implies that the performance

(32)

31 of peers, id est social performance feedback, can be a better indicator of required performance to survive.

However, post hoc analyses including environmental dynamism as a moderator on the following relationships showed to be insignificant: 1. Negative historical performance

feedback and research and development intensity, 2. Negative social performance feedback and research and development intensity, 3. Negative historical performance feedback and acquisitions, and 4. Negative social performance feedback and acquisitions. The results of the panel regression analyses are reported in the Appendix.

Dynamism is measured through regressing industry value of shipments over the five preceding years against time (1980–1984). Then the standard error of the regression

coefficient related to a time dummy variable is divided by the average value of industry's shipments to produce a standardized index of environmental dynamism (Boyd, 1995). The industries were defined using 4–digit SIC codes.

Analyses

To test the hypotheses, STATA was used to run the models. When running the analyses, the cross–section and time–series nature of the panel data were taken into account. Research and Development Intensity – Hypotheses 1 & 3.

To check for individual specific effects in the panel data, the Breusch–Pagan Lagrange multiplier test was conducted (Breusch & Pagan, 1980). The test rejected the null hypothesis of zero variance across units (X (4) = 4263.72, p < 0.001), which indicates significant

differences across firms in the data. Therefore, a simple OLS estimation is not appropriate for this study. In hypotheses 1 and 3, research & development intensity is the dependent variable of which the data has a range restriction from 0 to 1. This indicates the need for a panel Tobit regression model. In STATA, the command for a panel Tobit regression implies a random

(33)

32 effects model. However, there is no command for a parametric conditional fixed–effects model, as there does not exist a sufficient statistic allowing the fixed effects to be conditioned out of the likelihood. Consequently, I ran the Hausman test to determine whether to use the within (fixed effects) or Tobit (random effects) estimator. The results indicated the use of a fixed effects model, to account for firm and time effects (X (4) = 74.28, p < 0.001).

Therefore, the within (fixed effects) panel regression model was conducted to test hypotheses 1 and 3.

Acquisition Counts – Hypotheses 2 & 4.

In hypotheses 2 and 4, the dependent variable is acquisition counts, which is a count variable. Therefore, a simple OLS estimation is not appropriate for this study and this indicates the need for Poisson regression analyses or negative binomial regression analyses (Cameron & Trivedi, 2013).Poisson regression analysis assumes the mean and variance of acquisition counts to be the same. The mean value for acquisition counts is .977, yet the variance is 10.48, which indicates overdispersion in the data (see Table 1). Therefore, the negative binomial regression analysis seems more appropriate. However, the case of

overdispersion is rejected by the value of the Pearson dispersion statistic, as it is below 1 (1/df Pearson = .9776). This indicates that a within (fixed effects) Poisson panel regression model fits best to test hypotheses 2 and 4 (Wooldridge, 2010).

Results

The following section reports the results of this study. Firstly, the descriptive statistics and correlations between variables are discussed to provide an overview of the used sample. Then, the output of the multiple regression analysis is discussed in light of the first two hypotheses. Finally, the analyses and outcomes of the last two hypotheses are discussed.

(34)

33 Descriptive Statistics and Correlations

The relationships between the control, independent and dependent variables were examined using a Pearson correlation, excluding missing cases list wise. Table 1 reports the number of observations, means, standard deviations and correlation coefficients among the measures. Firstly, the correlations between the control variables and dependent variables are outlined. Then the correlations between the independent variables and the dependent variables are presented. Finally, some of the correlations between the control variables and independent variables are discussed. All correlations mentioned are significant at the p < 0.05 level.

When considering the relationship between research & development intensity (RDI) and the control variables, the results show that firms size correlates moderately negative with RDI (r=–.220). There was a moderately positive correlation between organizational slack and RDI, which was equal for absorbed slack (r = .171) and unabsorbed slack (r = .171).

Furthermore, the results indicate no correlation between RDI and distance from bankruptcy, as well as RDI and environmental dynamism. When considering the relationship between acquisition counts and the control variables, the results show opposite directions for all correlations compared to RDI. Firms size correlates moderately positive with acquisition counts (r=.315). There was a slight positive correlation between organizational slack and acquisition counts, which was slightly stronger for unabsorbed slack (r = –.111) than for absorbed slack (r = –.071).

Negative historical performance feedback correlates moderately negative to RDI (r = – .187), whereas positive performance feedback has a positive correlation with RDI (r = –.153). Conversely, negative nor positive historical performance feedback correlate to acquisition counts. RDI relates moderately negative to negative social performance feedback (r = –.278), as well as to positive performance feedback (r = –.180). Conversely, acquisition counts correlate weakly positive with positive social performance feedback (r = .045) and with

(35)

34 negative social performance feedback (r = .080). Negative historical performance feedback and negative social performance feedback are highly correlated (r = .877), which implies these variables cannot be taken together as independent variables in one model, due to multicollinearity concerns. Therefore, separate models for performance below historical aspiration level and performance below social aspiration level were conducted on the dependent variables.

(36)

TABLE I. Descriptive Statistics and Bivariate Correlations

N M SD 1 2 3 4 5 6 7 8 9 10

1. Research & Development Intensity 4509 .127 .118 –

2. Acquisition Counts 4509 .977 3.237 .006 –

3. Logged Firm Size 3892 –.453 2.009 –.220* .315* –

4. Absorbed Slack 3961 .511 .582 .176* –.071* –.214* –

5. Unabsorbed Slack 3961 3.649 3.181 .176* –.111* –.272* .453* –

6. Distance from Bankruptcy 3918 6.265 13.109 .031 .030 –.048* .224* .478* –

7. Environmental Dynamism 2505 .028 .017 .003 .013 –.125* .004 .037 .036 –

8. Performance below Historical Aspiration 3539 –.089 .25 –.187* .027 .159* .040 .085* .242* .000 –

9. Performance above Historical Aspiration 3539 .064 .173 .153* –.039 –.206* .009 –.048* –.106* .048* .133* –

10. Performance below Social Aspiration 3972 –.069 .255 –.278* .045* .206* .046* .127* .287* .014 .877* .022 –

11. Performance above Social Aspiration 3972 .146 .12 –.180* .080* .180* .006 .147* .326* –.002 .344* .136* .338*

Note: Acquisition Counts is a count-variable.

(37)

When interpreting the correlations between the performance variables and dependent variables, caution is needed. A negative correlation to performance feedback below aspiration levels indicates that the further performance falls below aspiration level, the more the variable of comparison rises. On the contrary, a positive correlation to performance feedback below aspiration levels indicates that the more performance falls below aspiration level, the more the variable of comparison falls.

Regression Analysis – Performance Feedback and Research and Development Intensity Table 2 presents a summary of the estimation results for the control and independent variables on research & development intensity. Model 1 shows a base model consisting only of the control variables and research & development intensity. The independent variables of historical performance feedback are added in model 2 and the independent variables of social performance feedback in model 3. Hypothesis 1 postulates that the effect of performance below aspiration on research & development intensity is positive. Model 2 and 3 show the coefficients are significant and negative for the main effect of performance below aspiration level for both historical (Model 2: t = –2.58, p = 0.010) and social (Model 3: t = –3.20, p = 0.001). This suggests that research & development intensity increases when performance drops further below aspiration levels. Hence hypothesis 1 is confirmed.

(38)

37 Hypothesis 3 suggests that the negative relationship between performance below historical aspiration level and research & development intensity is more negative than the relationship between performance below social aspiration level and research & development intensity. However, the high correlation between performance below historical aspiration level and performance below social aspiration level, indicates multicollinearity risks.

Therefore, separate models for performance below historical aspiration level and performance below social aspiration level were conducted on research and development intensity. To compare the regression coefficients between performance below historical aspiration level of Model 2, and performance below social aspiration level of Model 3, the following formula was used, where SEβ is the standard error of β (Clogg, Petkova, & Haritou, 1995):

Z = β2 − β3

(SEβ2)² + (SEβ3)²= 0.353

For the coefficients to differ significantly from each other, the Z–score is required to be greater than an absolute value of 1.64, where the α = 0.05. The found Z–score is 0.353, TABLE 2.

Within (fixed effects) Panel Regression Model for Research and Development Intensity

on person-organization fit and organizational commitment. Model 1 Model 2 Model 3

Variables β SE β β SE β β SE β

Logged Firm Size .007* .002 .004 .003 .009* .002

Absorbed Slack .010* .003 .003 .003 .010* .003

Unabsorbed Slack .006* .001 .008* .001 .007* .001

Distance to Bankruptcy –.000 .000 .000 .000 .000 .000

Environmental Dynamism .072 .089 .056 .091 .062 .088

Performance below Historical Aspiration –.018* .007

Performance above Historical Aspiration –.031* .011

Performance below Social Aspiration –.021* .007

Performance above Social Aspiration –.044* .013

Model F 21.30* 16.57* 19.46*

R .05 .06 .06

N 2403 2123 2403

(39)

38 which fails to reject the null hypothesis of similarity of the coefficients for the negative

performance variables in Model 2 and Model 3. This suggests that the negative relationship between performance below historical aspiration level and research & development intensity is not found to be more negative than the relationship between performance below social aspiration level and research & development intensity. Hence, hypothesis 3 is not supported.

Regression Analysis – Performance Feedback and Acquisition Counts

Table 3 presents a summary of the estimation results for the control and independent variables on acquisition counts. Model 4 shows a base model consisting only of the control variables and acquisition counts. The independent variables of historical performance

feedback are added in model 5 and the independent variables of social performance feedback in model 6. Hypothesis 2 postulates that the effect of performance below aspiration on acquisition counts is positive. Model 5 and 6 show the coefficients are positive, yet not significant for the main effect of performance below aspiration level for both historical (Model 5: z = –.96, p = 0.335) and social (Model 6: z = 0.20, p = 0.845). This suggests that acquisition activities do not decrease when performance drops further below aspiration levels. Hence hypothesis 2 is not supported.

Hypothesis 4 suggests that the positive relationship between performance below social aspiration level and acquisition activities is less positive than the relationship between

performance below historical aspiration level and acquisition activities. However, this hypothesis builds on hypothesis 2, which was not confirmed. Therefore, hypothesis 4 is not supported either.

(40)

39

Discussion

The following section reflects upon the theorized predictions and the consequent outcomes of this research. This study was set out to examine the relationship between negative performance feedback and organizational risk–taking behavior in the light of corporate entrepreneurship. The results of the data analyses are roughly in line with the predictions made. Furthermore, the role of the type of aspiration level on this relationship was considered, however, the analyses did not yield significant differences. Firstly, the major findings are reviewed. Subsequently, the contributions of this research are presented. Finally, the limitations of this study and the recommendations for future research are discussed.

Major Findings

There is substantial debate in the literature on the relationship between negative performance feedback and risk–taking behavior of top–level managers in organizations, due TABLE 3.

Within (fixed effects) Poisson Panel Regression Model for Acquisitions Counts

on person-organization fit and organizational commitment. Model 4 Model 5 Model 6

Variables β SE β β SE β β SE β

Logged Firm Size .996* .054 .959* .055 .996* .054

Absorbed Slack .007 .175 .083 .131 .012 .173

Unabsorbed Slack –.205* .037 –203* .033 –.207* .037

Distance to Bankruptcy –.006 .005 –.004 .005 –.008 .005

Environmental Dynamism 4.597* 1.766 3.956* 1.769 4.756* 1.78

Performance below Historical Aspiration –.122 .127

Performance above Historical Aspiration .242 .253

Performance below Social Aspiration .025 .127

Performance above Social Aspiration .183 .243

Model X² 552.14* 480.07* 551.14*

N 1517 1393 1517

Historical versus social performance feedback and risk-taking in corporate entrepreneurship : research & developement versus acquisitions