Performance feedback and organizational risk-taking behavior : the moderating effects of heterogeneity in top management team demographics

(1)

Thesis

Performance Feedback and Organizational Risk-Taking Behavior: The

Moderating Effects of Heterogeneity in Top Management Team demographics

Date of Submission: 22 June 2018 Course: Thesis

Course Lecturer: Bernardo Lima

(2)

Statement of originality

This document is written by Student Jordi Heere 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.

(3)

Abstract

This paper contributes to the behavioral theory of the firm performance feedback literature by addressing the gap in the literature considering low performance feedback within organizations and their risk-taking behavior. This research reconciles the gap by incorporating the interaction effects of performance feedback and Top Management Team characteristics (TMT). I posit that organizations respond differently to low performance feedback for different levels of TMT characteristics and heterogeneity within those characteristics. The characteristics are gender, age and tenure. I posit that females in the TMT are more risk-averse compared to men facing low performance feedback. For the characteristic age, I posit that older TMT members take less risk compared to younger members facing low performance feedback. For the characteristic tenure, I posit that TMT members with longer organizational tenure become more inert facing low performance feedback. Besides these the heterogeneity within TMTs in terms of these characteristics was also tested. I posit that a higher diversity within the TMT in terms of age results in more risk-taking behavior and for gender and tenure I posit that more diversity in the TMT would lead to a decrease in risk-taking behavior. I find that organizations with TMTs whose members are older respond more risk-averse towards low performance feedback compared to younger members. Partly are these findings in line with previous research and TMT members influence the response to performance feedback and organizational risk-taking behavior.

Key Words: Behavioral theory of the firm; performance feedback; aspiration levels; top

management team; demographic characteristics; upper echelons theory; heterogeneity top management team

(6)

Introduction

In the past four decades, a significant body of literature was developed about the ways firms learn and gain experience from performance feedback (Argote & Greve, 2007). Performance feedback and its effect on decision outcome is central in the behavioral theory of firms (March & Shapira, 1987). Most scholars have used behavioral theory of firms as a basis to research performance feedback (Argote & Greve, 2007). The perspective of the behavioral theory of firms suggests that decision-makers use aspiration levels to evaluate an organization’s performance (Cyert & March, 1963). The performance relative to the aspiration levels influences the risk-taking behavior of organizations (March & Shapira, 1987). Performance below the aspiration level indicates that the practices of an organization are not functioning as they are supposed to. Therefore, decision-makers are looking for solutions, based on performance feedback, to improve the situation (Cyert & March, 1963). Performance that lies above the aspiration level increases the decision-makers’ need to stay in the status-quo, because change is not necessary since the company is functioning as it is supposed to do at a satisficing level (Greve, 2003).

Research on how organizations try to improve when facing low performance is, however, contradictory. On one hand, decision-makers in organizations that face performance below their aspiration level, increase their risk-taking behavior because they think that changing their behavior can help to improve their performance (Singh 1986; Bromiley 1991; Wiseman and Bromiley 1996; Greve 1998). On the other hand, there are many scholars that argue differently, because there are contradicting results on the effect of low performance feedback on organizational risk-taking behavior (Argot & Greve, 2007). There are studies that argue that when a firm’s performance is below their aspiration level, they become rigid and do not

(7)

undertake any risky decisions, those studies contend that low performance leads to less risk-taking behavior (Miller & Chen, 2004; Audia & Greve, 2006).

A significant amount of studies on risk-taking behavior within organizations focused on conditions within organizations that are critical in shaping the risk-taking behavior within organizations (Audia & Greve, 2006; Desai, 2008; Lim & McCann, 2013). Recent research acknowledges the importance of including organizational conditions due to the fact that these influence organizations in distinctive ways, which leads to different forms of risk-taking behavior (Lim & McCann, 2013). Therefore, these conditions are essential in order to understand the relationship between performance feedback and risk-taking behavior (Audia & Greve, 2006; Desai, 2008; Lim et al., 2013).

The study of Desai (2008) focuses on the importance of researching experience, legitimacy and age as moderating factors on the relationship between performance feedback and risk-taking behavior. The results show that organizations with less experience in operating have less buffer from failure. Other results from the research of Desai (2008) shows that age-related inertia might decrease responsiveness to performance shortfalls under certain circumstances, however, these are not defined. Lim & McCann (2013) argue that when performance falls below aspiration levels, high stock compensation will increase risk aversion of CEOs. However, it will increase risk-taking propensity of outside directors, as their risk preferences are more in line with the organization’s shareholders. Therefore, it is necessary to research organizational conditions in the relationship between performance feedback and organizational risk-taking behavior.

In the relationship between performance feedback and organizational risk-taking behavior one important factor has not been researched; many researchers acknowledge that decision-making is influenced by the perception of risk (Keyes, 1985; Bromiley and Curley, 1992; Krueger and Dickson, 1994). The perceptions, however, of every decision-maker are

(8)

different. This difference in perspective, according to Wiersema and Bantel (1992), is based on the different values and experiences of top managers. Although characteristics of decision makers have great influence on their risk perceptions and risk-taking behavior, this has not yet been researched in this context.

According to Alessandri and Pattit (2014), it is important to look beyond just the individual in this research and take the entire top management team into consideration. As Top management team (TMT) characteristics have been proven to have a significant effect on the perceptions managers have, the perceptions and interpretations of the TMT critically influence strategic decisions (Dutton & Duncan, 1987). A TMT’s decision to initiate change in strategy will be based on each member’s perception of opportunities and constraints (Tushman & Romanelli, 1985; Figner & Weber, 2011). The upper echelons theory from Hambrick & Mason (1984) states that organization’s strategic directions can be explained by the demographic backgrounds of top management team members. As a TMT engages in the strategic decision-making process, each managers’ perceptions and interpretations will reflect their own cognitive base and values. Therefore, firms with distinctive top management teams interpret and respond to performance feedback differently when facing low performance feedback levels (Lim & McCann, 2013).

This study attempts to clarify the contradicting results on the relationship between performance feedback and organizational risk-taking behavior with the addition of taking differences between managers into account. In order to do this, this study scrutinizes the relationship between performance feedback and organizational risk-taking behavior for firms that perform below their aspiration level with the moderating effect of top management team characteristics. I build upon theory about risk-taking behavior and top management team characteristics as possible moderators, which are age, tenure and gender. With these characteristics, I propose that differences in the characteristics influence the risk preferences of

(9)

an organization, and therefore its propensity to take risks. These characteristics are used because TMT literature has identified several characteristics that have a proven effect on risk-taking behavior and are therefore, important to research within this context (Hambrick et al., 1984; Klenke, 2003).

According to Wiersema & Bantel (1992) more heterogeneous teams, in terms of educational level, and with younger team members are more willing to take risks. Organizational tenure was also found to have a positive effect on organizational change and risk-taking behavior of the TMT. Hambrick, Cho and Chen (1996) found that heterogeneity within teams increased action propensity and enhanced willingness to take risks. Homogeneity means that organizations are similar in their characteristics, for instance, demographic characteristics. Additionally, according to Hambrick et al. (1996) they were faster in executing actions and these were more in line with the status quo of the organization. Therefore, the effect of performance feedback on risk-taking behavior will be moderated by the degree of heterogeneity within top management teams, therefore it will be scrutinized as a boundary condition in this study.

In order to contribute to the debate of risk-taking behavior when a firm is functioning below their aspiration level, I try to contribute by adding the moderating effect of TMT characteristics as a boundary condition, because the TMT characteristics tenure, age and gender are characteristics proven to have a significant influence on risk preferences of managers within an organization (Finkelstein & Hambrick, 1990).

This paper contributes to the behavioral theory of the firm and literature about top management team characteristics through analyzing its moderating effects of TMT characteristics on the relationship between performance feedback and organizational risk-taking behavior. With this research, I try to provide a better understanding of the effects of TMT characteristics on the relationship between performance feedback and organizational outcome.

(10)

The following section describes the current body of literature and how this study will complement to it.

Literature Review

The following chapter discusses the current body of literature of performance feedback and heterogeneity of top management team characteristics, and how it influences risk-taking behavior of organizations. It is followed by an explanation of the literature gap, research question and hypotheses. The first section outlines performance feedback and the basis theory of its origin, followed by an explanation of its relationship with risk-taking behavior of organizations. Then, the literature gap is described and the research question is formulated. Lastly, the behavioral theory of the firm is linked with a proxy for risk-taking behavior and Top Management Team characteristics are introduced that will function as boundary condition.

Performance Feedback and the Behavioral Theory of the Firm

Performance feedback has its origins in the behavioral theory of the firm and is essential to the literature as it has been widely adopted (Cyert & March, 1963). The behavioral theory of the firm’s main assumption is that, as decision-makers are rationally bounded, they do not make perfect decisions regarding their surroundings. The decision-makers can only make decisions based on their perspectives and learn from their experiences and performance feedback. They base their decision on aspirations, by comparing their perceptions with those of their surroundings (Cyert & March, 1963). The behavioral theory of a firm explains that decision-makers use aspiration levels to evaluate an organization’s performance. The performance relative to the aspiration levels influences their learning and risk-taking behavior in order to respond to the situation (March & Shapira, 1987; Miller & Chen, 2004). Organizations use

(11)

aspiration levels to establish their organizational goals and by setting benchmarks they compare to their performance with other market players (Cyert & March, 1963). Aspiration levels are a minimum level of performance that organizations deem as satisfactory, and make up whether the organization is functioning successfully or not (March & Shapira, 1987). According to Cyert & March (1963) organizations compare themselves by focusing on two internal points of view. One point of view is that an organization compares itself to its past performance. The organization uses historical performance as a benchmark and tries to perform better in the future. This way of benchmarking the organization is useful because it evaluates the effectiveness of the organization to adapt to prior performance levels (Levinthal & March, 1981).

Another type of comparison of aspiration levels for an organization is to compare itself with industry peers. In this case it uses “social comparison” as a benchmark at which the company wants to perform minimally (Cyert & March, 1963). These targets are the minimum level at which a company wants to function and would be perceived as satisfactory. Performance below these levels are seen as not satisfactory and explained as routines or practices that do not meet the requirements of the environment (Desai, 2008). Social and historical aspiration levels are used to simplify the decision-making process because it is easier to interpret positive or negative performance instead of the continuous measure of performance (Chen, 2008).

These ideas about social and historical aspirations levels are, according to Audia and Greve (2006), based on psychological processes explained by Kahneman and Tversky (1996). These psychological processes include risk perception and risk preferences. This means that the organization takes action based on their perception of the performance feedback relative to the social or historical aspiration level. The psychological processes, therefore, influence organizational learning behavior and risk-taking behavior (Audia & Greve, 2006; Desai, 2008;

(12)

March & Shapira, 1987; Miller & Chen, 2004). Essentially, it is about the organization’s learning capacity, which is based on experience and the discrepancy between aspirations and achieved performance. This performance feedback provides organizations with information about how they are functioning and if they have to change their operations, search for alternatives or maintain their position (Chen, 2008).

According to the behavioral theory of the firm, when organizations function below their benchmark or aspiration levels, they tend to respond by taking more risk. This is because they want to resolve the problem, and thus they engage in search behavior in order to get back to the desired performance levels (Cyert & March, 1963). There are two types of search behavior defined by Cyert and March (1963). First, problematic search behavior, which is explained as search behavior that has an effect which is triggered when organizations fail to perform at their aspiration levels. The outcome is deemed as not satisfactory by the decision maker and organizations will search immediately for solutions for their low performance in order to return to their aspiration levels (Chen, 2008). The other type of search behavior is slack search. Slack search is stimulated by the organization’s slack resources, such as, excess resources (Levinthal & March, 1981). Slack facilitates investments that could help the organization with future changes in the environment and help to cope with those changes by adaptation possibilities (Chen, 2008).

When organizations function above their benchmark or historical or social aspirations, they tend to respond more risk-averse and inert, because organizations that are functioning well have no need to change their activities. Also, because engaging in more risky activities could have a negative effect on firm performance (Bromiley & Miller & Rau, 2001; Miller & Chen, 2004).

There are contradicting results in literature about search behavior of organizations. According to Audia and Greve (2006), stimulates a negative attainment discrepancy context

(13)

organizations to search for a solution. It predicts risk-taking behavior in the sense that the desire to overcome performance failure is larger than the desire to extend success (Kahneman & Tversky, 1996; Audia & Greve, 2006). Decision-makers respond by searching for solutions that could only address the performance shortfalls they are facing (Greve, 1998). Chen and Miller (2007) argue that low performance feedback leads to more risk-taking behavior in order to get back to performance levels that are deemed as satisfactory. Also, because organizations see the necessity to change in order to perform more towards their aspiration level (Singh 1986, Bromiley 1991, Wiseman and Bromiley 1996, Greve 1998, Palmer and Wiseman 1999; Iyer & Miller, 2008). Moreover, firms performing above their aspiration levels tend to become more rigid, inert, take less risk, and forego changes available to competitors, because they have less urge to change since the organization is functioning well (Bromiley et al., 2001; Greve, 2002; Miller & Chen, 2004).

On the other hand, several scholars argue that low performance feedback leads to less risk-taking behavior because firms restrict their activities to their core-business (Sitkin and Weingart 1995; Wiseman and Gomez-Mejia 1998). According to Miller and Chen (2004), this threat-rigidity effect is based on psychological responses to threat. When this happens in organizations, it triggers several actions such as: less information processing, tighter control, and less spending of resources. For example, when firms are close to bankruptcy they reduce risk (Miller & Chen, 2004). Performance feedback above and below the benchmarked aspiration levels have an effect on risk-taking behavior within organizations (Greve, 2003). Scholars on performance feedback above and below the aspiration levels have researched both effects and found various results (Lim & McCann, 2013; Desai, 2008; Audia & Greve, 2006). These scholars have based their research on various boundary conditions but still there is a lack of understanding on the matter. The results are inconclusive and inconsistent.

(14)

Scholars have responded to these differences by looking for the boundary conditions and test the boundary conditions to see whether low performance feedback leads to increased or less risk-taking behavior (Desai, 2008; Lim & McCann, 2013). By doing so they treated organizations as heterogeneous, because firms respond differently if they have different organizational conditions (Lim & McCann, 2013). Even though differences between organizations are highly relevant in explaining interpretations of performance feedback and different organizational risk-taking behavior. Meaning that organizations respond differently to different organizational conditions and, therefore, it is important to research (Lim & McCann, 2013). It is necessary to understand those boundary conditions and research contextual factors as they influence the effect of performance feedback. Conditions that are critical in shaping the risk-taking behavior of organizations (Desai, 2008; Lim & McCann, 2013).

For example, Lim and McCann (2013) showed that the amount of stock options influenced the risk-taking behavior of organizations when they faced low performance feedback. Studies that highlight the importance of heterogeneity between firms showed the importance of understanding influencing factors in organizational conditions. Those factors are essential in order to understand the relationship between performance feedback and organizational risk-taking behavior (Audia & Greve, 2006; Desai, 2008; Lim et al., 2013).

The study of Desai (2008) explains the debate as an experience differential between organizations and relates it to the age of those organizations. Older organizations become more inert and resistant to change compared to organizations that are younger. Because when organizations get older, they become part of developed networks of partners and invest in technologies that are more stable. Additionally, Desai (2008) argues that age-related inertia may decrease responsiveness to performance shortfalls under some circumstances, however, these are not defined.

(15)

Audia & Greve (2006) argued that the conflicting findings are caused by the heterogeneity and proposed several options that could explain the differences in the results. They found that low performance in small firms decreases organizational risk-taking, because small organizations do not have the resource buffer that larger organizations have, which leads to a different aspiration point. Smaller firms tend to focus more on aspiration closer to their minimal survival level, therefore, they react to low performance feedback with less risk-taking responses. These are all firm level characteristics except for the compensation moderator of Lim et al. (2013). This moderator has been researched extensively, but in a different context. According to Wiseman and Gomez (2007) stock option compensation leads to less risk-taking behavior because the decision-maker is personally more affected when a high risk-reward strategy fails. Scholars have tried to contribute to the debate by examining those boundary conditions when low performance feedback leads to risk taking or risk aversion. They tried to see organizations not as unitary agents, but rather as heterogeneous, as firms in different situations are likely to process information differently (Lim & McCann, 2013). Therefore, it is necessary to include organizational conditions when researching the effect of performance feedback on organizational risk-taking behavior.

Literature Gap and Research Question

The behavioral theory of the firm often used performance feedback to explain organizational risk-taking behavior because the effect of performance feedback above or below aspiration levels has been proven to significantly influence organizational risk-taking behavior (Bromiley et al., 2001; Audia & Greve, 2006; Desai, 2008). However, there are two sides to the matter. On the one hand, according a large body of literature, performance feedback below aspirations levels leads to less risk-taking behavior within organizations (Cyert & March, 1963; Bromiley et al., 2001). These scholars argue that these organizations become more rigid and inert because they allocate their resources to parts of the organizations with low risk (Audia &

(16)

Greve, 2006). The results are not consistent regarding the effect of performance feedback below the aspiration levels. On the other hand, another stream of literature argues that low performance leads to more risk-taking behavior, because those organizations are willing to take more risk in order to get back to satisfactory levels relative to their aspiration levels (Audia & Greve, 2006; Miller & Chen, 2004).

However, scholars have not researched individual and team characteristics when considering this relationship, even though people and groups are essential in decision-making (Cyert & March, 1963). The psychological responses of managers within organizations are based on the perspectives managers have and this influences decision-making (Hambrick & Mason, 1984). The “Upper Echelons Theory” describes that an organization’s performance is associated with Top Management Team characteristics. This theory states that organizational outcomes are a reflection of the characteristics that Top Management Team members possess. Decisions that managers make are the result of behavioral factors, and not so much of rational thinking (Cyert & March, 1963). Therefore, in in order to understand organizational behavior, it is necessary to identify the factors that influence or direct the attention of TMT members (Finkelstein & Hambrick, 1990).

Overall, researchers agree and acknowledge that decision-making is influenced by the perception of risk, and therefore organizations performance feedback differently (Keyes, 1985; Bromiley and Curley, 1992; Krueger and Dickson, 1994). March and Shapira (1987) suggest that when the level of perceived risk increases, a person is less likely to engage in risk-taking behavior because boundary conditions of the organizations influence the way people perceive risk. But according to Kahneman and Tversky (1979) the response of decision-makers is different and they do not engage in taking behavior, but rather they become more risk-averse when they face low performance feedback. According to Wiersema and Bantel (1992), this difference in response is based on the different perspectives managers have. The different

(17)

perspectives that decision-makers have, are in turn influenced by their demographic characteristics. Therefore, it is necessary to understand the demographic traits that influence the perspectives of top management team members in order to understand the decisions managers make when they face low performance feedback. Demographic traits are used because they function as a proxy for predictors of beliefs, values and abilities of managers, which influence risk-taking behavior (Hambrick & Mason, 1984).

These articles clearly demonstrate that it is necessary to further research the relationship between performance feedback and organizational risk-taking behavior and include individual and team aspects. Literature based on the upper echelons theory state that the individual cognitive base of the upper echelons in the organization come from experiences and are therefore an important boundary condition for risk-taking behavior. Therefore, it is important to include the upper echelons theory in this research, because it could explain differences in results from previous research. Hence, to obtain a better understanding of the relationship between performance feedback and organizational risk-taking behavior, it is important to consider the top management team demographic characteristics (and its composition) that influences this relationship. It is reasonable to expect that the firm’s top management team influences the organizational risk-taking behavior in a negative attainment discrepancy context. This research tries to give an insight into how TMT members moderate the effect of performance feedback on organizational risk-taking behavior. Therefore, I have set up the following research question that answers this gap:

What is the moderating effect of TMT characteristics and its composition on risk-taking behavior in a negative attainment discrepancy context, expressed in research and development expenditures?

(18)

Theoretical Framework

The following section describes the theoretical base necessary for this research. From the theoretical base, several hypotheses have been drawn up in order to answer the research question. In order to find the source of the differences established in previous research, the moderator and control variables will provide more insight into the relationship of performance feedback and organizational risk-taking behavior. These ideas are researched when organizations face negative attainment discrepancy context. The section starts with the explanation of the specific demographic traits effects and followed by heterogeneity effects on performance feedback and organizational risk-taking behavior. The concepts consist of TMT members age, TMT members’ organizational tenure and TMT gender composition and it is hypothesized that these concepts have an interaction effect on performance feedback and organizational risk-taking behavior additionally heterogeneity is an interaction variable that influences that relationship.

The upper echelons theory is an extension of the behavioral theory of the firm by Hambrick and Mason (1984), in their view the organization is a reflection of its TMT members is a better way to predict organizational outcomes compared to individual characteristics. Alessandri and Pattit (2014) built on this and state that it is important to look beyond just the individual and take the entire top management team into consideration, because the entire TMT is a better predictor for organizational outcomes. The TMT is the entity through which decisions are made. Those decisions depend on the TMT’s way of interpreting information regarding their environment where it recognizes problems and opportunities and adjusts the organizations to those stimuli (Hambrick & mason, 1996)

An examination of what influences executives and how executives assess situations and determine a firm’s strategy is an important area of investigation (Wiersema & Bantel, 1992).

(19)

This is especially the case as all top management team members play an important role in meeting the firm’s objectives and exert a direct influence on corporate decisions (Alessandri & Pattit, 2014). This influence is of importance because the strategic decision-making process as a whole is very complex and ambiguous. The top management team is defined as those who are in the upper tier of an organization (Hambrick & Mason, 1984). Theoretical interest in top management teams began together with the research about firm behavior (Cyert & March, 1963). According to this theory, decision-makers are often unable to make rational decisions because they are rationally bounded and they therefore act based on their interpretation or perception of their surroundings. The perceptions and interpretations of the TMT critically influence strategic decisions (Dutton & Duncan, 1987).

The current management literature on strategic choice and strategic management emphasize the importance of the TMT in the formulation and implementation of an organization’s strategy. As a TMT engages in the strategic decision-making process, each managers’ perceptions and interpretations will reflect their own cognitive base. Hambrick and Mason (1984) define the cognitive base as assumptions about future events, knowledge of alternatives, and the consequences attached to alternatives. They proposed a model of how a manager’s cognitive base influences the underlying process of decision-making. The cognitive base first limits the perception or places to which attention is directed. Second, because of selective perception, the manager only pays attention to a selective part of his/her environment. Finally, the processed information goes through the cognitive base of experiences and specific knowledge of the manager. When the decision-maker uses his/her cognitive base and values to make a decision, there is a gap between his/her perception and the real situation. This gap can be explained by a sequential process of perceptions. “The managers’ field of vision” is limited because a manager cannot scan his or her entire surroundings before making a decision. Secondly, because of selective perceptions, managers are biased in what they perceive and use.

(20)

Thirdly, the selected information is interpreted through a filter which consists out of the cognitive base and values of the managers. Therefore, a TMT’s decision to initiate change in strategy will be based on their member’s perceptions of opportunities and constraints, which is influenced by the performance feedback they perceive (Tushman & Romanelli, 1985; Figner & Weber, 2011). Those perceptions are influenced by the values and beliefs of managers. Demographic characteristics are indicators of quality (Hambrick & Mason, 1984). Demographic characteristics are often used in research as a proxy for beliefs, values and abilities. Consequently, several scholars have found a significant association between demographic composition of the top management team and organizational behavior (Hambrick & Cho & Chen, 1994). They acknowledge that human filter and social biases at the top of organizations influence the competitive behaviors of organizations.

The upper echelons theory from Hambrick and Mason (1984) continue on this matter and state that organization’s strategic directions can be explained by the demographic backgrounds of top management team members. The research of Hambrick et al. (1984) demonstrated that there are several demographic characteristics that have a significant effect on decision-outcomes of decision-makers. These characteristics are: educational level, functional background, experience, age, gender and tenure. Many scholars have found significant effects between these demographic characteristics and organizational characteristics. Some have found that organizations with younger and short tenured TMT members had a positive effect on innovation, where long organizational tenure of a TMT tenure was found to have a strong association with absence of change (Hambrick & Mason, 1996; Wiersema & Bantel, 1992). The demographic variables scrutinized in this paper are age, tenure, and gender respectively. These demographic characteristics are interesting to research, as they suggest receptivity to change and willingness to take risk (Wiersema & Bantel, 1992). Additionally, these characteristics are interesting to research because these characteristics have been proven to

(21)

affect the cognitive base and information processing abilities of TMT members (Finkelstein & Hambrick, 1990; Katz, 1982).

The composition of those demographic TMT characteristics is also interesting to research for several reasons. According to Wiersema & Bantel (1992) it does matter whether the TMT consists out of a heterogeneous team or a homogeneous team. Having Heterogeneity within a TMT characteristic gives the team diverse perspectives and interpretation. Katz (1982) argues that a high team diversity enables the TMT to be more dynamic and more able to adapt to changing environments. Other benefits of having a diverse group composition is higher levels of creativity and innovation and this leads to more novel ideas (Wiersema & Bantel, 1992). The driving logic behind this idea is that diversity increases the variety of perspectives, cognitive resources and overall problem-solving capabilities of a group (Hambrick & Mason, 1996). Due to their diverse backgrounds and orientations, heterogeneous TMT’s are able to observe more opportunities, threats and several fronts. This enables the TMT to a larger potential of reacting to changes (Hambrick & Mason, 1996). Moreover, this heterogeneity is based on differences within cognitive bases and values of top managers and how they interpret and process information. According to Hambrick & Mason (1984), a heterogeneous TMT collects their information based from different sources and will have different perspectives. The benefit of heterogeneous groups is that different viewpoints lead to novelty, diversity and comprehensiveness in the set of recommended solutions. TMT’s with this characteristic will be able to challenge each other’s viewpoints (Hoffman & Maier, 1961). However, at high levels of diversity the communication within a TMT will become more difficult and conflict loaded, which can lead to the inability to take action (Wiersema & Bantel, 1992). However, Wiersema and Bantel (1992) argue that high levels of diversity within a TMT leads to the inability of making decisions as communication becomes more difficult. The diversity can lead to internal conflicts that could slow down the decision-making process. Having a more homogeneous TMT

(22)

will make it easier for a team to reach a consensus because of shared values and beliefs as wel as because of conformity. The process of information is influenced by this (Wiersema & Bantel, 1992). Therefore, the levels of heterogeneity, besides the demographic characteristics, are necessary to research because they could influence the relationship of performance feedback on organizational risk-taking behavior.

The effect of homogeneity with respect to demographic concepts was shown by Janis (1972), in which it was explained that homogeneity leads to cohesiveness and inferior decision-making. He states that homogeneity brings on groupthink and this results in fewer idea generation and alternatives for decision-making. Homogeneity means that organizations are similar in their demographic characteristics. Compared to heterogeneity, which means that organizations are different in their characteristics. TMT homogeneity resulted in lack of openness, and exhibited conformity (Hambrick & Mason, 1984). On the other hand, TMT homogeneity was also found to be useful for decision-making processes where interdependence characterizes a firm’s diversification posture, which demands a high level of integration among TMT members (Wiersema & Bantel, 1992). Also, homogeneous teams were found to be useful for more routine problem-solving tasks where a diversity of perspectives is not necessary (Filley & House & Kerr, 1976). TMT’s that have a more homogeneous or heterogeneous composition will, therefore, respond differently to certain factors. These papers suggest that the reaction of TMT members to low performance is moderated based on the team composition and that they take different decisions concerning risk-taking behavior because of the difference in the TMT composition. Therefore, when organizations face a negative attainment discrepancy context, heterogeneity will positively moderate the relationship of performance feedback and risk-taking behavior. This way the use of the Upper Echelons Theory in addition to the performance feedback theory can give new insights into the matter and explain more of the variance in results in previous research.

(23)

Organizational TMT Age and Heterogeneity

The age of members within a TMT is a characteristic of TMT members that has proven to have influence on the cognitive base of managers (Hambrick & Mason, 1984; Wiersema & Bantel, 1992). Age is an accurate indication of cognitive frameworks and skills of managers. TMT members’ age has been found to have a negative relation to the ability to integrate new information and to make risky decisions (Wiersema & Bantel, 1992). Therefore, a TMT member’s age influences strategic decision-making. TMT members that are older tend to seek more information before they make decisions. Older managers need more time to evaluate the information. Both result in more time that TMT members need before they make decisions (Hambrick & Mason, 1984). Besides the extra information and the more time TMT members need before making decisions, security of income and career is more important to older TMT members, which leads to more risk-averse decisions (Wiersema & Bantel, 1992). This means that an increase in age will lead to an increase in time necessary to evaluate performance feedback, and the quality of processing the feedback reduces, because of decreasing cognitive ability, memory, learning ability, and reasoning (Bantel & Jackson, 1989).

Younger TMT members tend to be more risk oriented and their age has been associated with organizational growth and volatility (Wiersema & Bantel, 1992). According to Chen & Huang (2010), younger managers are better able to integrate information and learn. Moreover, their education is more recent and their financial security and career are less of an issue at an early stage. And with their ability to process and integrate information quicker than older TMT members, they seize opportunities quicker (Wiersema & Bantel, 1992). It has been proven that, as younger TMT members process information less carefully, they therefore make different decisions regarding risk-taking behavior in a negative attainment discrepancy context. They engage quicker in problematic search behavior and risk-taking behavior, and by doing so, when

(24)

they face low performance feedback, they will take more risk than other, older TMT members. Therefore, in the examination of performance feedback and organizational risk-taking behavior, TMT member’s age will negatively moderate this relationship, as TMT members will differently interpret performance feedback and differently respond to it.

H1a. Higher levels of TMT age will negatively moderate the relationship between low performance feedback and organizational risk-taking behavior versus lower levels of TMT age, resulting in less R&D expenditures when performance decreases.

As stated in the literature review, heterogeneity within TMT characteristics influences the decision-making behavior of TMT’s. A higher age diversity within TMT’s will lead to less group think and more novel ideas with an increase in innovation expenditures as a response to low performance feedback (Wiersema & Bantel, 1989). The composition of the TMT varies more with TMT members that have different experiences. Similar age will lead to similar experiences of environmental events, which in turn leads to shared values and beliefs. A heterogeneous team in term of age will lead to an increase of different perspectives and less group think. People are less inclined to change their behavior to the other group members because of less identification. These differences through age within the TMT lead to a higher consideration of risk which has been proven to lead to higher levels of innovation (Wiersema & Bantel, 1992). Therefore, I posit that heterogeneity within a TMT in terms of age positively moderates low performance feedback on organizational risk-taking behavior.

H1b. Higher levels of Age heterogeneity within a TMT will positively moderate the relationship between low performance feedback and organizational risk-taking behavior versus

(25)

lower levels of heterogeneity, resulting in more R&D expenditures when performance decreases.

Organizational TMT Tenure and Heterogeneity

There are not many TMT characteristics studied sufficiently, but organizational tenure of TMT has been researched quite extensively compared to other demographic characteristics within this context (Pfeffer, 1983). Tenure is the duration that a TMT member has spent within an organization to the measurement date. Differences in TMT tenure are expected to explain why some firms react on performance below aspiration level with risk-taking and other with risk-aversion (Finkelstein & Hambrick, 1990). Therefore, organizational TMT tenure can contribute to the ongoing debate in performance feedback literature, because it is acknowledged to influence the cognitive base and subsequently, information processing and decision-making (Finkelstein & Hambrick, 1990). Organizations where TMT members are groomed within the organization, and therefore, have a long organizational tenure, tend to conformity with the status quo within their organization (Hambrick & Mason 1984). Katz (1982) also argued that TMT tenure is associated with increased rigidity and commitment to established policies and practices. Long tenure also has other implications on a different level. TMT members in organizations who are for a long time within said organization have spent years to get to their position. This means that they have a lot more to lose and are therefore more risk-averse because they have less to gain with risky activities (Katz, 1982). Strategic conformity is another effect of long TMT tenure within organizations. The longer TMT members stay in organizations, the more they reduce the adoption of new strategies and bring the organization to an industry conformity. As individuals spent more time in an organization they become committed to established policies and practices

(26)

In contrast, TMT members with shorter tenure are more willing to take risks. They have diverse information and their perspectives are less restricted to a status quo of the organization or industry (Wiersema & Bantel, 1992). Short-tenured TMT members who take more risky decisions are expected to have very high returns or very low returns because of their riskier endeavors. When tenure increases the firm will get closer to industry performance averages. Longer-tenured TMT members change less and imitate more in a risk-averse manner (Wiersema & Bantel, 1992). There are some studies, however, that argue that long-tenured TMT members are more likely to take risks because their experience results in having less uncertainty, but according to the upper echelons theory it is acknowledged that an increase in TMT members tenure leads to conformity to a situation of status quo. Therefore, I posit that TMT members with short organizational tenure are expected to engage in more problematic search and reach towards a risk-taking manner when facing low performance feedback.

H2a. Higher levels of TMT Tenure will positively moderate the relationship between low performance feedback and organizational risk-taking behavior versus lower levels of TMT tenure, resulting in more R&D expenditures when performance decreases.

Organizational tenure of a TMT member indicates experiences of the respective TMT member, which produces similar sorts of interpretations that TMT members have towards their environment compared with other members of the TMT that have a similar tenure length within the organization (Wiersema & Bantel, 1992). Homogeneity in terms of organizational tenure leads to commitment to the status quo and group confirmation. And these two factors lead to risk-aversion, which means that the organization will not increase their risk-taking behavior when facing low performance feedback (Wiersema & Bantel, 1992). A heterogeneous TMT in terms of organizational tenure has different experiences within the organization and this results

(27)

into less conformity because the members have joined the organization in different stages. Their response to low performance feedback will be different as well, since their experiences within the organization are different. A more diverse board has the advantage of having the experience of longer tenured TMT members but also the TMT members that bring a new perspective to the organization. And a higher level of diversity leads to an increased desire and ability for change and risk-taking behavior (Li & Wahid, 2017). For these reasons I posit that, heterogeneity in terms of tenure within a TMT negatively related to low performance feedback and organizational risk-taking behavior.

H2b. Higher levels of Tenure heterogeneity within a TMT will negatively moderate the relationship between low performance feedback and organizational risk-taking behavior versus lower levels of heterogeneity, resulting in more R&D expenditures when performance decreases.

Organizational TMT gender composition and Heterogeneity

The gender composition within a top management team is also acknowledged to influence the way managers process information and make decisions (Johnson & Powel, 1994). Literature within this field demonstrates that women are more risk-averse than men (Jianakoplos & Bernasek, 1998). These conclusions are based on research that showed that female TMT members experience less volatile earnings and have lower leverage, which resulted in a higher survival rate compared to companies with more male TMT members (Faccio et al, 2014). Currently, it is generally acknowledged that women are more conservative and risk averse than men. This results in less taking within organizations and less risk-taking behavior for personal gain (Jianakoplos & Bernasek, 1998). Women are more committed to stability in combination with female TMT members being more risk sensitive. When

(28)

performance is low they accept lower returns (Johnson & Powel, 1994; Fehr-Duda et al., 2006). There is a difference in interpretation between female and male executives (Johnson & Powel, 1994). Male TMT members are less risk-averse and often engage in more risky activities (Jianakoplos & Bernasek, 1998). They tolerate a higher level of risk and volatility when facing low performance in order to gain a higher pay-off and return to a satisfactory performance level (Johnson & Powell, 1994). Therefore, having more male TMT members within an organization will result in more risk-taking in a negative attainment discrepancy context. Consequently, I posit that men positively moderates the effect of low performance feedback on organizational risk-taking behavior.

H3a. Higher levels of females within the TMT will positively moderate the relationship between low performance feedback and organizational risk-taking behavior versus lower levels of TMT age, resulting in less R&D expenditures when performance decreases.

The composition of gender within TMT explains interpersonal barriers to reach agreements on strategies (Richard et al., 2004). TMT’s that are more homogeneous in terms of gender have less of those barriers and this makes it easier to overcome those barriers and reach agreements. However, groups that have division in the TMT that includes similar amounts of men as women leads to an increase of group formation within the TMT. The TMT members may identify themselves stronger with their own gender, which could lead to a decrease in the quality of intergroup communication and the inability to act when facing changes in the organization or environment like low performance feedback (Richard et al., 2004). And TMT that consist of more women tend to take less risk than TMT with more men when facing low performance feedback (Wahid, 2012). Therefore, I posit that heterogeneity within the TMT in

(29)

terms of gender negatively moderates the effect of low performance feedback on organizational risk-taking behavior.

H3b. Higher levels of Gender heterogeneity will negatively moderate the relationship between low performance feedback and organizational risk-taking behavior versus lower levels of heterogeneity, resulting in more R&D expenditures when performance decreases.

Methodology

This section will explain in more detail the methodology of this research. The section starts with a description of the overall design, followed by the used sources, then the sample description and subsequently the used databases and how they were operationalized. Consequently, the variables are explained, and we describe what alterations were necessary with the addition of a description of the used models to analyze the data.

Sample

The sample used in this research is the US manufacturing industry, as the intended way to measure the dependent variable of risk-taking behavior is via research and development intensity. Following prior research, research and development intensity is best measured in

(30)

manufacturing organizations (Lim & McCann, 2013). The firms within the US manufacturing industry are labelled with Standard Industrial Classification (SIC) codes. The numbers of the manufacturing lie within the range of 2000 and 3999 and only these 4-digit codes are used. Within these codes this research will only use firms within the S&P500 to gain more complete data from 2010 till 2014. S&P500 firms are used because they need to publish more data about their firm by law. This will help with the completeness of the data and ease to compare. This data is found via the Wharton Research Data Services website. In order to attain the data, the systems used are Compustat and Execucomp (Iyer & Miller, 2008; Wiseman & Bromiley, 1996). Compustat has annual and quarterly financial data for companies in the U.S. who are on the stock exchange. These include companies who are currently in the stock exchange and also companies who were in the stock exchange. Data from Execucomp has quarterly and annual data, which includes relevant information about top management teams.

The data retrieved from Compustat and Execucomp were merged based on two criteria. The first criterion is a company identifier GVKEY: the GVKEY uniquely identifies companies, which is necessary in order to match the datasets. The second criterion is the fiscal year FYEAR. This makes it possible to combine the company data also on the correct date. For all the firms in this dataset, information regarding their top management teams will also include their demographic characteristics specified above. The top management team is specified as members of the board of directors. Because the dataset with the top management team variables, contained several lines of data per year, Stata made duplicates for every variable in the dataset with the other variables in order for them to match correctly. These duplicates were dropped after calculating the right variables per GVKEY per FYEAR for the moderators, which is explained in the sections ‘moderators’ in more detail.

(31)

Dependent Variable

The dependent variable in this study is organizational risk-taking behavior measured by R&D intensity. Following prior research, this study will measure this variable by using the research and development intensity. This way, the relative size of the firm is also taken into account (Chen & Miller, 2007; Lim & McCann, 2013). Research and development is risky because it implies a large investment that consists out of a lot of uncertainty and the potential of significant losses (Palmer & Wiseman, 1999). R&D investments require a large investment in a short period of time and the payoffs may come at a much later moment (Lim & McCann, 2013). Hence, R&D intensity is a good proxy for organizational risk-taking behavior because high investment in R&D is often a strategy for high returns (Lim & McCann, 2013).

The mean of R&D intensity is 0.079 and the standard deviation is 0.104. The median is 0.044 and the skewness is 3.509.

Independent variable

The independent variable in this study is firm performance. Following prior research, Firm performance will be measured with the return on assets (ROA) (Bromiley, 1991; Desai, 2008; Greve, 2003; Lim & McCann, 2013). It is measured by comparing the ROA with the social and historical aspirations of companies. In order to measure these two concepts, two variables were made. The Social aspiration variable, that compares the company to industry peers, is measured by calculating the mean ROA of all the companies within the manufacturing industry (4 digit SIC level). The historical aspiration variable, that compares the company to its past performance, is measured by calculating the firm’s average return on assets over the past three years. Performance below aspiration levels had only observations that were less than zero and the observations above zero were above aspiration levels compared to their past performance or compared to the industry peers. Following Desai (2008), this was measured by

(32)

splitting the variables into two categories. Performance below aspiration level is 0 when performance is above the aspiration level. Performance above the aspiration level is 0 when performance is below the aspiration level (Audia & Greve, 2006). In order to test the relationship of performance feedback on organizational risk-taking are the independent performance variables lagged by one year. The aspiration levels were at t-2 years. Following prior research, all the performance variables are lagged one year compared to the dependent variable (Lim & McCann, 2013).

Moderators

This section first explains how the moderators of the demographic characteristics were operationalized and followed by an explanation of their heterogeneity measurement. The dataset containing the dependent, independent and control variables were data per GVKEY per FYEAR, the data regarding the moderating variables (retrieved from Execucomp) contained more data-lines per GVKEY and FYEAR, because there were several TMT members. Therefore, several adjustments were necessary to make in order to calculate the interaction variables. Following Wiersema & Bantel (1992), the moderating variables were TMT member age, organizational TMT tenure and the TMT gender composition calculated by taking the average of all the top management team members per company per fiscal year. The top management team’s tenure was the total amount of years a manager has spent within the organization. It was used by taking the starting date minus the fiscal year it was focused on. The average tenure was used per company per year for the entire TMT in order to calculate the interaction variable (Finkelstein & Hambrick, 1990). Stata, however, uses a different measurement, edates, for time variables and the variables were made continuous before subtracting the starting date with the specific FYEAR. Subsequently, the number of days were formatted into a number of years. For the calculation of top management team’s gender

(33)

“0” and a female “1”. The part of male within the TMT was divided by the entire number of executives within the TMT. This resulted in a proportion of females within the TMT. From the proportion of females, the average proportion of females was calculated per company per year. The dataset, however, contained as said above several lines per GVKEY per FYEAR. So, the averages were calculated and the duplicates were dropped. For the calculation of the heterogeneity variables the coefficient of variation was used in the dataset before the average calculation and the duplicates drop. The coefficient of variation is a scale efficient measure and, therefore, preferred to other measures like just the standard deviation. It is measured by dividing the standard deviation by the mean of the variable (Finkelstein & Hambrick, 1996). The heterogeneity variable TMT members age, organizational TMT tenure and organizational TMT gender composition were calculated as said by taking the coefficient of variation per GVKEY per FYEAR.

From all the moderating variables, the interaction variables were made for the regressions by generating new variables. These were calculated by multiplying the moderating variables with the independent variables creating interaction variables. This happened for all the TMT characteristic variables and the heterogeneity variables in combination with the low performance feedback variables with social and historical aspiration levels.

Control Variables

Several variables have been proven to influence the relationship of performance feedback and organizational risk-taking behavior and are therefore included as control variables in this research. Previous research of March and Shapira (1982) showed that firms close to survival were more risk-averse. Because of this reason, organizational survival will be included as a control variable. This is measured with the Altman Z-score. Based on previous research the Altman z-score is operationalized (1,2 x ((Current Assets – Current Liabilities)/ Total

(34)

Assets)) + (1,4 x (Retained Earnings/ Total Assets)) + (3,3 x (EBIT/ Total Assets)) + (0,6 x (Market Value of Equity/ Total Liabilities)) + (Sales/ Total Assets).

Furthermore, larger firms have greater resources, but are also often more bureaucratic and this effect of firm size was proven to have a significant effect in a negative attainment discrepancy context and therefore will also be included as a control variable (Audia & Greve, 2006). Firm size will be measured as the logarithm of the number of employees within the organization (Audia & Greve, 2006).

Also, following prior research of Chen and Miller (2007), the effects of slack are included in this research. Three types of slack are taken into account. Potential slack, unabsorbed slack and absorbed slack. This was measured by dividing the selling, general and administrative expenses by the sales of the company for the absorbed slack. Potential slack was used by using the debt to equity ratio. The current ratio was used as a proxy for available slack by subtracting the current liabilities of the company from the current assets (Chen & Miller, 2007)

Statistical Model

This research used a panel data set and the cross-sectional, time-series character of the data were taken into account for the decision of picking the right models. The downside of panel date, higher levels of heteroskedacity, was corrected for by using a robust model in the regressions. The heteroskedacity was tested with the Wald test in a fixed effect regression model( (76)= 2.3e+34, p < 0.05). For testing for multicollinearity, the uncentered variance inflation factors was used because the centered may fail to discover collinearity involving a constant term. In this way, the constant is viewed as a legitimate explanatory variable in the regression. The mean VIF is 4.10, which means that with one outlier of firm size, meaning that there are more predictors. However, in this research this is not interesting and, therefore,

(35)

was not necessary in this dataset since the dataset is too short, because it only consists of four years. After considering the individual effects test, the Hausman test was used to determine whether the data needs to be handled with a fixed-effect model or a random-effect model, caution was necessary because it tested one-sided. The Hausman test showed significant results ( (7)= 32.23, p < 0.05), which indicates that the fixed effects model is appropriate for running the regressions and interaction effects for testing the hypotheses. Because it only tested one-sided, time effects were also taken into account. After examining the data, outliers were removed based on the 2%-98% trim Windsor method. Furthermore, missing values of moderators were also removed. However, after correcting the data the heterogeneity variable, TMT tenure resulted in only a few observations and was not complete enough to be useful in this research and was therefore excluded.

Results

This section describes the results of this research. It will start with the descriptive statistics of the variables used in this research to give an overview of the data. Followed by a correlation analysis that shows a table that reports the correlations of the used variables. Lastly, the regressions were carried out with interaction effects in order to test the hypotheses from the theoretical framework.

Descriptive Statistics and Correlation Analysis

This section first describes basic statistics of the dataset after correction and also the bivariate list wise correlation analysis between the dependent, independent, moderating and control variables. Table 1 below shows the outcomes of the summary of the dataset and the correlation analysis. After merging the datasets and manipulating it as described in the previous section, the end number of observations is 480 from 2010 until 2014. The organizations within

(36)

this dataset within the US manufacturing industry are all with the SIC codes between 2000 and 3999 that are or were in the S&P 500.

Looking at the correlations in table 1, the performance variables above and below the aspiration levels were significantly correlated to R&D intensity. Performance feedback below historical aspiration levels was significantly negative correlated to R&D intensity (r=-0.609). This also holds up for performance feedback below social aspiration levels (r=-0.044). Performance feedback above the historical aspiration level was positively correlated to R&D intensity (r=0.223). Performance feedback above the social aspiration level was also positively correlated to R&D intensity (r=0.189). Considering the moderating variables, not all of them were significantly correlated to R&D intensity or the independent variables. R&D intensity and TMT gender was significantly negative correlated (r=-0.126). This is also the case for the TMT age (r=-0.108) and TMT tenure (r=-0.084). Social TMT age is also significantly negatively correlated to R&D intensity (r=-0.068). The heterogeneity age variable historical is significantly negative correlated to R&D intensity (r=-0.095). This also holds up for heterogeneity within gender social (r=-0.053) and age social (r=-0.010). When looking at the moderators and the independent variables, historical TMT gender is significantly positive correlated to performance below aspiration level (r=0.277). Historical TMT age is also significantly positive correlated to historical low performance feedback (r=0.304). This is the same for historical TMT tenure (r=0.274). The heterogeneity variable TMT gender is also significantly positive correlated to performance feedback above historical aspiration levels (r=0.294). Noteworthy, however, is that these four moderating variables are not significant correlated to performance feedback above historical aspiration level or performance below social aspiration levels. They are negatively correlated to performance feedback above social aspiration levels. For TMT gender, historical is it 0.400). For TMT age, historical is it

(37)

all significantly negative correlated; Age 0.352), Gender 0.230) and Tenure

(r=-0.404). When looking at the dependent variable and control variables it shows that, survival is

positively correlated to R&D intensity (r=0.098). Firm size is significantly negative correlated to R&D intensity (r=-0.377). Absorbed slack is significantly positive correlated to R&D intensity (r=0.793) and this is the same for available slack (r=0.329). When considering the control variables and the independent variables, firm size is significantly positive correlated to performance feedback below historical aspirations (r=0.306). Absorbed slack is significantly negative correlated to performance feedback below historical aspirations (r=-0.442). Absorbed slack is significantly positive correlated to performance feedback above historical aspirations (r=0.235) and also with performance feedback above social aspirations (r=0.290). The other correlations are shown in table 1 below.

(38)

(39)

Regression Analysis

The results of the regression analysis were split up into table 2, historical aspiration levels and table 3, social aspiration levels. For both aspiration levels the same regressions were executed in 5 different models. The 5 models were executed consequently and were appended and divided into the two tables. The significant results are highlighted with *** (p<0.01), **(p<0.05) and *(p<0.10). For table 2 and 3, model 1 only contains the dependent variable, R&D Intensity, and the control variables. In model 2 in table 2 and 3, the independent performance feedback variables are appended, above and below aspiration levels. Starting from model 3, in table 2 and 3 the panel time series effects are included to test its effect on the control and independent variables. Model 4 in table 2 and 3 adds the interaction effects of the demographic character interactions; TMT member’s age, TMT member’s tenure and TMT member’s gender. The regressions in model 5 include also the heterogeneity interaction variables; TMT age composition, TMT gender composition. These interactions were created based on the relationship between the dependent variables R&D intensity and the respective independent variable. In table 2, this is made for historical aspiration levels and in table 3 this is made for social aspiration levels.

The results in the tables need to be interpreted as following: positive coefficients positively influence R&D intensity for its respective independent variable, historical or social, and negative coefficients do the opposite. Meaning that if an organization has low performance feedback and the coefficient is positive, the lower the historical performance goes below the aspiration level, the lower the R&D intensity is. For a negative coefficient that means, in the same context, the R&D intensity will increase.

In table 2, the first model only includes the control variables and the coefficient of determination is very low (R2 = 0.068), meaning that the control variables only partly explain the variation in the model. By building in the independent performance feedback variables, the

Performance feedback and organizational risk-taking behavior : the moderating effects of heterogeneity in top management team demographics

Thesis