The moderating role of environmental dynamism on the relation between negative performance feedback and organizational risk behavior

The moderating Role of Environmental Dynamism on the

Relation between Negative Performance Feedback and

Organizational Risk Behavior

Master Thesis of Caroline Reinecke University of Amsterdam

Thesis topic: Performance Feedback

Supervisor: Dhr. MSc. B. Silveira Barbosa Correia-Lima

University of Amsterdam, Amsterdam Business School

Student: Caroline Reinecke (11153296)

MSc. Business Administration – Strategy

Statement of Originality

This document is written by Caroline Reinecke who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Table of Contents

Abstract………..……….3

Introduction………..…………..……….4

Literature Review………..…………..………....7

Theoretical background on learning from performance feedback and risk………8

Literature Gap and Research Question………..…………..………...10

Boundary Conditions………...11

Environmental Dynamism………..…………..…………...13

Low and High Risk………..…………..………...14

Hypotheses Development………..…………..………...16 Method………..…………..………..17 Data Collection………..…………..…………..…………18 Measures………..…………..………...…19 Statistical Model………..…………..………....…21 Results………..…………..……….……..17

Descriptive Statistics and Correlation Analyses…………..……….……22

Regression Analysis…………..…………..……….……..………….………..27

Discussion………..…………..……….……....30

Major Findings………..…………..……….……….31

Contributions………..…………..……….…………34

Limitations and Future Research………..…………..……….…….…35

Conclusions………..…………..……….………...37

References………..…………..……….………...38

Tables and Figures Table 1: Descriptive Statistics and bivariate correlations………...25

Table 2: Fixed effects panel regressions for high risk……….…...26

Table 3: Fixed effects panel regressions for low risk………..………...27

Figure 1: Moderating effect of environmental dynamism on high risk…………...30

Abstract

The behavioral theory of the firm and the prospect theory suggest that negative performance feedback increases organizational risk-taking. The threat-rigidity theory however proposes that organizations become more risk averse as performance falls below their aspirational levels. As both perspectives have found empirical support, boundary conditions need to be identified in order to specify conditions under which an organization becomes either more risk seeking or more risk averse as a response to negative performance feedback. High levels of environmental dynamism make it increasingly difficult for managers to foresee the future as dynamic environments are highly uncertain. This study investigates, whether the effect of negative performance feedback on organizational risk taking differs under different levels of environmental dynamism. Accordingly, it was hypothesized that firms become more risk averse whenever the environment is more dynamic. If the environment is more stable, firms were expected to react with increased risk-taking as a response to negative performance feedback. Furthermore, this study distinguishes between high and low levels of risk. Based on financial data of U.S. manufacturing firms, this study shows that firms choose safer options in the face of highly dynamic environmental conditions in order to offset the uncertainty of the environment. Whenever the environment is more stable, they actively search for ways to improve performance and accept higher levels of risk. These results were largely consistent with the prior hypotheses.

Keywords: organizational learning, organizational risk-taking, the behavioral theory of the firm, performance feedback, environmental dynamism, high and low risk

Introduction

As the fast paced business environment of the 21st century is very dynamic, businesses need to constantly renew themselves in order to ensure survival (Agarwal & Gort, 2002). Literature on organizational behavior largely agrees on the fact that organizations base their decisions regarding future courses of actions on performance feedback (Cyert & March, 1963). According to the behavioral theory of the firm, organizations use aspirations in order to evaluate their performance. This performance feedback then guides major strategic decisions concerning research and development (R&D) intensity (Greve, 2003; Chen & Miller, 2007), acquisitions (Iyer & Miller, 2008) and organizational change (Greve, 1998), making it a highly relevant research topic within the field of organizational behavior (Shinkle, 2011). An aspiration level is a desired performance outcome, which helps organizations to evaluate their current performance as success or failure (Shinkle, 2011). They are formed by comparing current performance to past performance or to performance of industry peers, and influence organizations inclination towards change and risk (Greve, 2003; Chen & Miller, 2007). The theory is based on psychological processes of risk perception and preferences (Kahneman & Tversky, 1979; Staw, Sandelands & Dutton, 1981) and organizational processes of search (Cyert & March, 1963).

The literature on organizational behavior further argues, that whenever a firm’s performance falls under the aspiration level, manager’s engage in more problemistic search in order to identify alternative courses of action that will help the firm perform better in the future (Cyert & March, 1963). Conversely, firms performing above their aspiration level and thus receiving positive performance feedback, engage in less problemistic search, trusting the effectiveness of their current routines (Cyert & March, 1963). Scholars of strategic management have taken the outcomes of performance feedback a step further and argue that they not only affect problemistic search, but also influence organizational risk preferences (e.g. Audia & Greve, 2006; McKinley, Latham & Braun, 2014; Bromiley, Miller & Rau,

2001; Desai, 2008; Greve, 2003; Miller & Chen, 2004; Nickel & Rodriguez, 2002). However, they report inconsistent findings regarding the direction of risk. While previous scholars largely agree on the fact that firms performing above the aspirational level engage in less risk-taking (Bromiley, Miller & Rau, 2001, Nickel & Rodriguez, 2002), contradictory views exist whenever it comes to firms performing below their aspiration level (McKinley, Latham & Braun, 2014, Audia & Greve, 2006). One group of researchers argues that firms become more risk seeking because the need for improvement becomes more salient in the face of organizational decline (Cyert & March, 1963; Kahneman & Tversky, 1979). The other line of the argumentation suggests that firms become more risk averse and rigid as negative performance stresses the presence of danger (Lopes, 1987; Sitkin & Pablo, 1992; Staw, Sandelands & Dutton, 1981). As research has yield inconsistent empirical results, this debate has not yet settled (Shinkle, 2011).

These inconsistent results indicate that performance shortfalls might be perceived differently under different conditions. Research on boundary conditions is needed in order to identify conditions under which performance below aspiration is perceived as a threat, leading organizations to become more risk averse or less as a threat, leading them to become more risk seeking (Audia & Greve, 2006). Such knowledge is crucial in order to gain a profound understanding of organizational risk-taking as a response to performance shortfalls. To date, research concerning boundary conditions has focused mainly on internal organizational factors such as resources, slack, size and experience (Miller & Leiblein, 1996; Audia & Greve, 2006; Shinkle, 2011), while external factors have been largely overlooked (Shinkle, 2011). However, as strategic decision-making is strongly influenced by the resources available to them and the predictability of future events (Aldrich & Pfeffer, 1976), including environmental conditions in this debate could enhance our understanding of organizational risk-taking. Environmental characteristics affect firm’s perceptions of risk in the sense that risk within an environment that is highly uncertain, might seem more threatening compared to

the same risk within an environment characterized by stability (Dess & Beard, 1984). Previous studies show, that the environment plays a role in firm’s aspirations (Levinthal & March, 1981) as well as in firm’s risk preferences (Baucus & Near, 1991; Mishina, Dykes, Block & Pollock, 2010). This indicates that it is relevant to include environmental factors when studying organizational risk-behavior as a response to performance shortfalls below aspirations. However, environmental factors have been largely overlooked in the debate of the effects of negative performance feedback, which is why this study takes a closer look at the role of the environment in the performance feedback – risk relationship. As environmental dynamism is the aspect of the environment, which is most strongly correlated to organizational outcomes (Dess & Beard, 1984), this present study will focus on the effects of environmental dynamism.

Next to environmental factors, attributes of the particular decision might also affect whether an organization feels threatened in the sense that the level of risk attached to different options might intensify the perception of threat. However, previous studies have largely overlooked that different actions in response to performance feedback might contain different levels of risk (Greve, 2003). This study attempts to advance research in this field by distinguishing between low and high levels of risk. Through adding this distinction, this study draws a more nuanced picture of organizational risk-taking but might also be able to give indications about how different levels of risk affect the organizations perception of risk-decisions.

This study attempts to reconcile contradictory findings of previous research through investigating the moderating effect of environmental dynamism in the relationship between negative performance feedback and risk-taking. The results of this study have two main theoretical contributions. First, findings enhance the behavioral theory of the firm by including environmental conditions in the research field. Through identifying the environment as a boundary condition, this study explains, why prior research has found inconsistent results

by specifying external contingencies under which declining performance below the aspiration level leads to either risk-seeking behavior or risk aversion. Second, this study aims to draw a more nuanced picture of organizational risk-taking by distinguishing between low and high risk. This further advances the understanding of organizational behavior when performance declines, as organizations might perceive future decisions differently depending on the level of risk they contain.

Theory and Literature Review

In the following paragraphs, the underlying theories on which main insights of this research area are based are explained and the existing literature on performance feedback is reviewed. Next, the research gap is specified. Subsequently, the relevance of studying boundary conditions is outlined, the role of environmental dynamism is elaborated and the importance of distinguishing between different levels of risk is explained. Then, the hypotheses are outlined, based on the before explained theories. Following, it is explained how this research will be conducted. Next, the results are reported and discussed in relation to the prior literature. Finally, a conclusion will be made.

Theoretical Background on Learning from Performance Feedback and Risk

Performance feedback is an important source of learning for organizations. It influences a wide range of future activities, organizational behavior and preferences (Audia & Greve, 2006; Desai, 2008; Jordan & Audia, 2012; Kim, Finkelstein & Haleblian, 2015; March and Shapira, 1987; Miller and Chen, 2004; Miller and Leiblein, 1996; Singh, 1986). Research on organizational risk-taking is grounded on three central theories: the behavioral theory of the firm (Cyert & March, 1963), prospect theory (Kahneman & Tversky, 1979) and the threat-rigidity theory (Staw et al., 1981). The behavioral theory of the firm is based on bounded rationality, choosing satisficing options instead of maximizing and sequential attention (Cyert

& March, 1963). This perspective argues, that managers act rational until a certain degree but in the end are constrained by bounded rationality, meaning they are not able to take all possible outcomes into account. Managers therefore use goals and aspirations to simplify the assessment of performance. The discrepancy between goals and actual performance results in performance feedback, which guides future actions (Cyert & March, 1963). Prospect theory focuses on the framing of decisions and recognizes that fundamentally different decisions are made in the face of losses and gains (Kahneman & Tversky, 1979). The two theories base their arguments on a few shared assumptions. Both agree, that strategic behavior is guided by the organizations aspirations. The behavioral theory of the firms distinguishes between two different sources of aspirations: historical aspirations and social aspirations. Performance feedback based on historical aspirations is derived from comparison of an organizations current performance with their own past performance, while performance feedback based on social aspirations is derived from comparison of organizations current performance with the performance of industry peers (Cyert & March, 1963). In Prospect theory, the aspiration level is a status quo or a value of zero (Kahneman & Tversky, 1979). Both theories assume that outcomes above their aspirations are perceived as successes while performance below their aspirations are perceived as failures. They argue that a firm’s desire to overcome times of failure is stronger than the goal to extend success in better times. Accordingly, they state that organizations will engage in more efforts to improve whenever their performance is below their aspirations and predict that managers increase their engagement in problemistic search in the face of organizational decline (Cyert & March, 1963).

Problemistic search refers to managers actively searching for alternative courses of action in order to achieve more favorable outcomes in the future (Cyert & March, 1963). Engagement in problemistic search is the dominant reaction to a firm facing performance shortfalls as they indicate that previous routines have not proved to be effective. When the organization fails to meet aspired performance levels, managers are primed with the presence

of loss and failure. The need for improvement becomes more salient, leading managers to assess a greater variety of options for future courses of actions (Cyert & March, 1963; Kahneman & Tversky, 1979). Conversely, organizations that exceed their aspirations engage in less problemistic search, as there is no need to change effective routines (Greve, 2003).

The threat-rigidity theory predicts a different effect. It argues, that declining performance can be seen as a threat, which makes the presence of loss and failure shift into the focus of managers. Accordingly, threat triggers anxiety, which lowers the ability to process information leading to a lower variety of options. Individuals restricted by anxiety are more likely to fall back on well-learned routines. At the group level, in-group/out-group discrimination will rise which will lead to a more centralized decision making process, including only few managers. Translated to the firm level, these processes will restrict communication, making elaborate discussions less likely. All these factors result in less flexibility to change behavior, ultimately leading to rigidity (Staw et al., 1981). Based on these three theories, two different lines of argumentation have emerged in the ongoing discussion about risk seeking following negative performance feedback.

While most scholars agree on the fact that organizations engage in less problemistic search whenever they exceed their aspirations (Bromiley et al., 2001; Greve, 2003), there is an unsettled debate about the effects of performance feedback below their aspirations. Based on arguments of the behavioral theory of the firm and the prospect theory, one group of scholars argues that organizations engage in more risk when performance feedback falls further below aspirational levels (Audia & Greve, 2006; Boyle & Shapira, 2012; Desai, 2008; Miller & Chen, 2004). The negative feedback indicates that there is a problem, which stimulates search for solutions (Cyert & March, 1963). Managers try to identify alternative courses of actions and pursue new activities, which might improve the performance of the firm and lead to performance outcomes that meet their aspirations (Greve, 2003). The other group of scholars however argues that a firms’ willingness to take risks decreases in the face

of organizational decline (Staw et al., 1981; Greve, 2003; March & Shapira, 1987). They suggest that organizations see performance feedback below their aspirational level as an essential threat. This perception will lead to high levels of anxiety and stress, which restricts information processing and leads to rigidity. Options that entail a notion of risk are not considered, which makes adaptation unlikely. The second argument states that risk behavior is a motivational predisposition, differing between individuals. Accordingly, most people try to avoid uncertainty and rather stay with an unsatisfactory but predictable state than risk additional losses (Kahneman & Tversky, 1979).

Literature Gap and Research Question

Previous studies on performance feedback theory concerning the effect of performance below aspirational levels on firms’ risk preferences has yield inconsistent empirical findings (e.g. Audia & Greve, 2006; Desai, 2008; Bromiley, 1991; Mone, McKinley & Barker, 1998), offering support for both views. Some studies have found evidence for the perspective that firms are more willing to pursue new activities and engage in greater risks in the face of organizational decline (Audia & Greve, 2006; Desai, 2008) while other scholars argue that performance below aspirational level leads firms to stick to their core businesses and become more risk averse (Audia & Greve, 2006; Bromiley, 1991; Mone, McKineley & Barker, 1998). These inconsistent findings indicate that the current stream of literature is lacking knowledge on the effects of performance feedback below aspirations. In order to reconcile these opposing predictions it is necessary to learn more about possible conditions that account for these differences (Audia & Greve, 2006). According to the shifting attention model (March & Shapira, 1992), manager’s decision making is highly dependent on their focus of attention, which can differ depending on particular internal and external characteristics of the organization (March & Shapira, 1992). Therefore, this present research

focuses on such characteristics which function as boundary conditions, explaining when negative performance feedback results in risk seeking behavior and when it will lead to rigidity. Past research has focused on internal organizational characteristics such as slack and size (Miller & Leiblein, 1996, Audia & Greve, 2006), while external characteristics have been largely overlooked. Yet, there are fundamental indicators that environmental characteristics influence risk preferences of organizations (Wiklund & Shephard, 2003; Mishina, Block & Pollock, 2010; Dess & Beard, 1984). This research therefore sheds more light on how external conditions might explain the contradicting findings on the effect of performance literature. Furthermore, previous research has overlooked that different actions in response to negative performance feedback contain different levels of risk. This restricts a profound understanding of organizational risk-taking as a response to performance feedback. The current study acknowledges such differences by distinguishing between low and high risk activities. Hence, this study attempts to answer the following research question:

What is the moderating effect of environmental dynamism on the relation between performance below aspirations and organizational risk-taking concerning decisions containing high and low degrees of risk?

Before diving into the effects of environmental dynamism and different levels of risk, this study tests the baseline model expecting that performance shortfalls below aspirations leads to increased engagement in risk, seeking to replicate findings based on the behavioral theory of the firm (Cyert & March, 1963) and the prospect theory (Kahneman & Tversky, 1979). This leads us to hypothesis 1:

H1: When performance is below the aspiration level, performance decreases lead to increased risk-taking.

Boundary Conditions and the Shifting-Attention Model of Risk Taking

According to the three theories outlined in the previous section, managers’ decisions concerning risk are guided by their aspirations (Cyert & March, 1963; Kahneman & Tversky, 1979). However, March & Shapira (1992) suggest that this view is incomplete by proposing the shifting-attention model of risk-taking. Based on extensive studies on how managers perceive risk (Shapira, 1989) they discovered that instead of focusing solely on the aspiration level, managers can also assess their performance in comparison to the survival point, which leads to different risk behavior. The survival point refers to performance, which is so low that the organization fails. Whenever organizations perform below their aspirations but above their survival point, they can either focus on the aspiration point or the survival point. Focusing on either of those points holds different implications regarding risk-taking behavior. Whenever managers focus on the aspiration point, they interpret performance shortfalls as repairable gaps and are willing to take increasing risks as a response to negative performance feedback in order to achieve a turnaround. If managers however focus on the survival point, performance shortfalls are perceived as a threat of future existence, which leads to high levels of anxiety and stress, restricting further information processing. This focus leads to a decrease in risks as a response to negative performance feedback. With this model, March & Shapira (1992) reconcile the inconsistent findings of previous studies suggesting that firms become either risk averse or risk seeking depending on their focus of attention. March and Shapira (1992) explain, that the focus of attention shifts according to various rules.

In order to advance the shifting-attention model of March and Shapira (1992), Audia and Greve (2006) proposed two such rules. First, they assumed that firms focus on the reference point that is closer to their actual current performance as it is more probable to reach that point. Second, they propose that this is influenced by their stock of resources. They conducted a study taking firm size as a proxy for resource stocks and confirmed that mangers of bigger firms focus on the aspiration point whereas managers of smaller firms focus on the

more proximate survival point (Audia & Greve, 2006). Accordingly, bigger firms perceive risk as less threatening because they have greater resource stocks to fall back on in case of failure. This makes risky options less dangerous and leads to bigger firms being more risk seeking. Smaller firms however have fewer resources and will go close to bankruptcy if the risky investigation does not hold the expected value. They therefore perceive risks as more dangerous and are less likely to engage in new activities (Greve & Audia, 2006).

Desai (2008) conducted a similar study, integrating research on organizational legitimacy into the field of organizational behavior. The study shows that firms with higher legitimacy and more operating experiences are more likely to interpret organizational decline as repairable gaps, leading them to engage in increased risk-taking. Just like Audia & Greve (2006) showed that bigger firms have greater resources stocks to fall back on, Desai (2008) shows that firms with high levels of legitimacy and more operational experience have a bigger buffer against organizational failure, leading them to feel less threatened when performance declines.

Such studies have leveraged valuable insights to the debate concerning the effect of negative performance feedback on organizational risk-taking and call for more studies, specifying rules about when organizations focus their attention on either the survival or the reference point. While most research on moderating variables have focused on internal organizational characteristics such as resources, size, slack and experience (Miller & Leiblein, 1996; Audia & Greve, 2006; Desai, 2008), relatively little attention has been paid to external factors concerning the environment in which firms operate (Shinkle, 2011). The facts that environmental factors influence organizational learning (Shinkle, 2011) as well as perceptions of risk (Near & Baucus, 1991), make it a highly promising research topic within the discussion about the effects of negative performance feedback on risk-taking (Shinkle, 2011). Especially environments, which are highly dynamic, characterized by instability and turbulence are known to be particularly hard for managers as they then need to assess a

greater variety of possible courses of actions while their outcome is difficult to predict (Eisenhardt & Bourgeois, 1988). Therefore, the present study researches the moderating effect of environmental dynamism on the relationship between negative performance feedback and organizational risk-taking and thus seeks to identify under which environmental conditions firms respond with risk-taking on performance feedback below aspirations and when they respond rather risk averse.

Environmental Dynamism

Environmental characteristics are a multi-level concept, which can be split in three dimensions: munificence, complexity and dynamism (Dess & Beard, 1984). Munificence is the extent to which the environment offers sufficient resources and growth opportunities (Starbuck, 1976). Munificence implies an environment, which is growing and allows the organization to generate more slack resources (Cyert & March, 1963). Environmental complexity refers to the heterogeneity and range of an organizations activity (Dess & Beard, 1984). The environment is called dynamic, whenever there is much change, which cannot be predicted well beforehand. It is a turbulent environment, which forces managers to process a greater amount of information in order to achieve a given level of performance (Galbraith, 1973). Keats and Hitt (1988) further developed the dimensions of the external environment and found that environmental dynamism is the dominant influence, responsible for the biggest effects on firms’ decisions and performance. It is considerably more challenging to manage a firm under a highly dynamic environment, as it is hard for decision makers to learn from performance feedback when cause-effect relationships are difficult to interpret (Priem, Rasheed & Kotulic, 1995). The value of routines, services and products might fluctuate and change fast. The outcome variety associated with risky changes is even greater under turbulent environmental conditions, which makes it hard for managers to discern information and predict patterns (Eisenhardt & Bourgeois, 1988). This leads to managers suffering from

greater information processing burdens and higher levels of anxiety and stress (Tushman, 1979).

Based on these characteristics, this present study proposes another rule, extending the shifting-focus model of attention of March and Shapira (1992): A highly dynamic environment makes managers attention shift to the survival point, leading them to prefer less risky decisions. Highly dynamic environments might trigger perceptions of danger in managers, as the outcome of future actions will be even more uncertain than they are under stable environmental conditions. Because of the high variability of possible outcomes of actions, a failure of such an activity might offset the organization into bankruptcy. Managers might try to mitigate the dynamism of the environment through decreasing their engagement in risk and reduce the variability of possible outcomes, leading to the next hypothesis:

H2: Environmental dynamism negatively moderates the relation between negative performance feedback and risk, turning the positive effect of negative performance feedback on risk negative.

As mentioned earlier in this study, actions as a response to performance feedback contain different levels of risk (Greve, 2003). As this has been largely overlooked in previous research (Schimmer & Brauer, 2012; Iyer & Miller, 2008; Desai, 2008; Chen & Miller, 2007; Lim & McCann, 2014), the following section will outline why it is crucial to acknowledge differences in risk and will further specify the above presented hypotheses.

Low and High Risk

The outcome of problemistic search is a set of different options for future courses of actions. These options contain different levels of risk. In classical decision theory, risk is defined as the variation in the distribution of possible outcomes following a particular choice

and their likelihood (Pratt, 1964; Arrow, 1965). Accordingly, activities are seen as highly risky when the variance of possible positive and negative outcomes is high. In practice however, manager’s perception of risk is mainly driven by the possibility of negative outcomes, called the downside risk (March & Shapira, 1987). This view defines risk as the increased likelihood that investments will not meet aimed success. In this research, risk will be conceptualized as downside risk because this definition is closer to people’s perceptions of risk in real life (March & Shapira, 1987; Cyert & March, 1963; Miller & Leiblein, 1996). The conceptualization of risk needs to be stressed in order to interpret the results accordingly.

Previous studies have mainly overlooked that different actions in response to performance feedback might contain different levels of risk (Schimmer & Brauer, 2012; Iyer & Miller, 2008; Desai, 2008; Chen & Miller, 2007; Lim & McCann, 2014). For instance, investing in activities that the organization has engaged in before and that have proofed effective in the past is considered less risky than engaging in activities that are new to the organization (Greve, 2011). Almost all organizations constantly invest in R&D as R&D investments are believed to be crucial in order to stay compatible (Greve, 2003). In that sense, R&D expenditures can be considered as containing lower levels of risk, as firms can draw on past experiences when engaging in such expenditures (Greve, 2011; Markovitch, Steckel & Yeung, 2005). High risk changes, on the other hand, are options that are associated with a high variability of outcomes that cannot be predicted easily (Dess & Beard, 1984). For instance, capital investments can be considered as highly risky expenditures as firms might not be able to predict their future value (Markovitch, Steckel & Yeung, 2005). For example, a manufacturing firm investing in a new factory, can be considered as an investment containing high levels of risk, as it is a long-term investment that is specialized to the organization and cannot be easily disposed if it turns out to be useless (e.g. when demand decreases) (Greve, 2011; Markovitch, Steckel & Yeung, 2005). Kacperczyk, Beckman and Moliterno (2015)

point out that not differentiating between high and low levels of risk limits the research within this field.

Drawing back on the shifting-attention model of March and Shapira (1992), another rule concerning different levels of risk can be specified. The level of risk might influence manager’s focus of attention and trigger different responses to negative performance feedback. High risk actions might lead managers to focus more on the survival point, as high-risk activities such as large capital investments are mostly irreversible. When they turn out to be not necessary it might threaten the survival of the organization and lead to bankruptcy. A failure of a low-risk action on the other hand might bring the organization further away from their reference point however, won’t threaten survival. Accordingly, whenever the future action is characterized by a low degree of risk, managers might focus their attention to the aspiration point and are willing to accept more risks. In this sense, a higher level of risk might lead managers to shift their attention to the survival point which influences their perception of risk. Therefore, a specification of the above named hypotheses is made, adding a distinction between high and low levels of risk. Therefore, the two above named hypotheses are further specified into “low risk” and “high risk”.

In the first hypothesis, it was expected that negative performance feedback leads to an increase in risk-taking. However, after introducing the argument that high levels of risk might shift the attention of managers towards the survival point, enhanced risk-taking is only expected as a response to negative performance feedback when the level of risk is considered low. High-risk activities are expected to be perceived as a possible threat to existence and are therefore more likely to decrease as a response to negative performance feedback.

H1a: When performance falls further below the aspiration level, performance decreases lead to less high risk.

H1b: When performance falls further below the aspiration level, performance decreases lead to more low risk.

Accordingly, the second hypothesis, which stated that environmental dynamism negatively moderates the relationship between negative performance feedback and risk, is also further specified in high risk and low risk. The negative effect of negative performance feedback on high risk is believed to become even more negative under highly dynamic conditions while the positive effect of negative performance feedback on low risk is believed to turn negative as high environmental dynamism is believed to shift the focus of managers towards the survival point.

H2a: Environmental dynamism moderates the relation between negative performance feedback and high risk, in the sense that it strengthens the negative effect of performance shortfalls on high risk.

H2b: Environmental dynamism moderates the relation between negative performance feedback and low risk, in the sense that it turns the positive effect of negative performance feedback on low risk negative.

Method

In this section, the methods of this research are specified. First, the most important aspects of the sampling method are summarized. Then, the proxies used to measure the variables are outlined. Finally, it is specified which model is used to analyze the data.

Data Collection

This study was conducted in a quantitative manner using secondary financial data from the COMPUSTAT – SDC Thompson merged database. The data was restricted to

manufacturing industries to allow comparison with similar previous studies and to prevent possible influences from the industry, which might affect data. The sample contained financial data of companies over a time span of 10 years. Furthermore, the sample was restricted to manufacturing industries with SIC codes from 2000 to 3999 in order to be able to compare the results with previous studies who also restricted their sample to manufacturing firms (Chen, 2008; Iyer & Miller, 2008, Lim & McCann, 2014). This restriction prevents possible influences from the industry, which might affect data. The manufacturing industry is further split up in product categories (e.g. food, chemicals, electronics). Industries containing less than five firms were excluded from the dataset to prevent biases during the regression analysis (Chen & Miller, 2007). Following Chen and Miller (2007), firms with R&D expenses greater than sales were excluded from the sample because this study is based on the behavioral theory of the firm by Cyert and March (1963) which argues, that firms learn from past performances and adjust their processes accordingly. This however is not applicable to firms who invest more than they sell. Therefore, the sample is restricted to firms with R&D intensity less than or equal to 1.

As outlined in the theoretical framework, performance feedback is drawn from two different aspirational levels: one based on past performance (historical aspiration) and one based on the performance of other firms (social aspirations) (Cyert & March, 1963). However, as these pose different measures and might yield different results (Greve, 2003; Chen & Miller, 2007), they were separated in the analysis to be able to see possible differences in the effect. Furthermore, it needs to be acknowledged that even though this study focuses on performance below aspirational level, performance above aspirations was included in the theoretical model in order to study a complete concept. The results and discussion however focus on performance below aspirations.

Dependent Variable

Low-Risk. Low risk-taking will be proxied with changes in R&D expenditures (Markovitch, Steckel & Yeung, 2005). According to Markovitch, Steckel & Yeung (2005) managers rate the risk level of R&D relatively low. Furthermore, such changes occur frequently than highly punctuated strategic decisions which is why their effects can be predicted relatively easily (Greve, 2003). This variable is computed as the growth rate of R&D investments from t – 1 to t.

High Risk. This study proxies high risk with capital expenditures (CAPEX) including all expenditures concerning assets that are newly purchased or investments that improve the useful life of an existing capital asset. This proxy is chosen, based on the study of Larraza-Kintana, Wiseman, Gomez-Mejia & Welbourne (2007) who asked 69 members of an executive MBA class to rate the risk level of six activities holding considerable risk. The results showed, that capital expenditures are considered to hold significantly more risk than R&D investments. Capital expenditures are a risky strategic decision because their consequences are uncertain. Investments are usually highly specific and depend partly on difficult to predict environmental factors (Audia & Greve, 2006; Greve, 2003; Henderson & Cool, 2003; Palmer & Wiseman, 1999). Because they are mostly specialized and immobile, they cannot be easily disposed whenever they turn out to be not needed (e.g. a new factory cannot be easily sold when demand decreases) (Audia & Greve, 2006). This variable is computed as the growth rate of capital expenditures from t – 1 to t.

Independent Variables

Performance Feedback. To measure performance feedback, this study takes ROA

(net income total assets) as a proxy for firm performance. This will be used because in the manufacturing industry this is the main measure for firm profitability (e.g. Audia & Greve, 2006; Bromiley, 1991; Chen & Miller, 2004; Desai, 2008; Greve, 2003; Iyer & Miller, 2008).

Next, two variables are created which offset (1) firm performance against their own past performance of the prior three years for historical aspirations and (2) against the average of industry peers for the social aspiration. A spline of these variables will distinguish performance below aspirations from performance above aspirations, resulting in four performance feedback variables.

Moderating Variables

Environmental Dynamism. As industry dynamism reflects the amount of change in

each industry (Dess & Beard, 1984), this variable will be calculated by repressing time against industry sales for the five years preceding data collection. Then, the standard error of the regression slope coefficient was divided by the mean sales value to obtain the value from dynamism (e.g. Lepak, Takeuchi & Snell, 2003; Dean & Snell, 1996; Keats & Hitt, 1988).

Control Variables

Firm Size. Considering that the size of the firm might increase firm risk-taking (Audia

& Greve, 2006), it is controlled for firm size. In this industry, most previous studies have taken the number of employees as a proxy for firm size (Audia & Greve, 2006).

Distance from Bankruptcy. As previous studies have found that firms close to

bankruptcy act more risk averse, I choose to control for distance from bankruptcy (March & Shapira, 1992). Following past studies, I measure distance from bankruptcy with the Altman’s Z-score as “(1,2 x working capital divided by total assets) + (1.4 x retained earnings divided by total assets) + (3.3 x income before interest expense and taxes divided by total assets) + (0.6 x market value of equity divided by total liability) + (1.0 x sales divided by total assets)”. The greater the outcome value, the lower the risk of bankruptcy (Chen & Miller, 2007; Iyer & Miller, 2008; Lim & McCann, 2014; Miller & Chen, 2004).

Industry R&D Growth. Because industry prospects could influence firm’s investment

decisions, the industry sales growth is included as a control variable (Chen & Miller, 2007). This variable is computed as the percentage of change in industry sales from t – 1 to t.

Slack. As prior research suggests that slack resources influence growth (Greve, 2003;

Singh, 1986) slack was included as a control variable. Following Chen and Miller (2007) and Singh (1986) the current ratio (current assets divided by current liabilities) and the working capital to sales ratio were used as proxies for available slack. First, they were normalized and then they were summed to form a composite slack index (Miller & Chen, 2007).

Statistical Model

In the choice of the statistical model, it was considered that the data is characterized as panel data. Analyses for panel data were chosen accordingly. As all variables are continuous, several regressions will be made in order to analyze the effects. I performed two Hausman’s tests, in order to test whether a fixed effects model or a random effects model is more appropriate. The Hausman’s tests rejected the null hypothesis for both low (X²(7)= 33.86, p<.01) and high risk (X²(7)= 120.91 , p<.01) as a dependent variable, indicating that a fixed effects model fits the data best. The fixed effects model controls for unobserved heterogeneity when heterogeneity is constant over time and correlated with independent variables (Yaffee, 2003). In order to mitigate any potential multicollinearity problems, the predictor, the moderator and the control variables were mean-centered before creating the interaction variables (Aiken & West, 1991). In order to test the hypotheses, stata statistical package was used to test the (fixed effects) model.

Results

In the following, the results of this study will be presented. Descriptive statistics and bivariate correlations can be found in table 1. I will briefly discuss relevant correlations.

Subsequently, the regression results will be presented. They will provide an answer to the research questions of this study, developed in the theoretical part of the paper.

Descriptive Statistics and Correlation Analyses

The financial data ranged from 1979 to 2014 and consisted of 40.713 firm-year observations. The major two-digit industry segments were electronic and electrical equipment except computer equipment (22,46% of the sample), measuring, analyzing and controlling instruments (18,28%), industrial and commercial machinery and computer equipment (18,25%) and chemicals and allied products (15,78%).

Table 1 shows the descriptive statistics of the sample as well as the bivariate correlations. First, the correlation between the dependent variables (high and low risk) and the control variables (industry R&D growth, firm size, distance to bankruptcy, absorbed and unabsorbed slack) will be presented. Next, the correlation of the dependent variables and the independent variables (performance feedback below historical aspirations and performance feedback below social aspirations) will be discussed. Finally, the correlations between the dependent variables and the moderator (environmental dynamism) will come into focus.

The correlations show, that R&D growth of the industry is positively correlated with low risk (r = 0.03, p < .01) and high risk (r = 0.01, p < .01). Unexpectedly, firm size is negatively correlated with high risk (r = - 0.04, p < .01) and low risk (r = - 0.05, p < .01). Similar, distance to bankruptcy is negatively correlated to high risk (r = - 0.02, p < .01) and low risk (r = - 0.01, p < .05) indicating that firms invest in less risk activities whenever they are moving further away from bankruptcy. Slack however is positively correlated to high risk (r = 0.02, p < .01) and low risk (r = 0.06, p < .01), indicating that firms engage in more risk activities as their availability of slack resources increases. Looking at the independent variables, it can be seen that performance below historical aspiration is negatively correlated to high risk (r = -0.01, p < .01) and positively correlated to low risk (r = 0.01, p < .01).

Performance below social aspirations is negatively correlated with high risk (r = - 0.01, p < .01) however not correlated with low risk (p = .45). However, it has to be acknowledged, that all correlations are very weak (r < 0.3). Seen that the correlation between performance below historical and social aspiration is comparatively high (r = .46, p > .01), separate models are used for the regression analyses (Chen & Miller, 2007; Iyer & Miller, 2008; Lim & McCann, 20014). According to Greve (2003), managers prefer social aspirations when they feel like their firm is comparable to other firms and rely more on historical aspirations whenever they feel their firm is unique. Separated models give room for such discrepancies and avoid redundant indicators that could distort parameters estimates (Gordon, 1968). Regarding the moderator, environmental dynamism is not correlated with high risk (p = .14) and weakly negatively correlated with low risk (r = - 0.01, p = .05) indicating that low risk-taking decreases, as the environment gets more dynamic.

Table 1 Descriptive Statistics and bivariate correlations

**. Correlation is significant at the 0.01 level *. Correlation is significant at the 0.05 level

Variables N Mean SD 1 2 3 4 5 6 7 8 9 10 1 High risk 47.648 0.66 14.7 0 - 2 Low risk 47.115 0.25 1.93 0.03** - 3 Industry R&D growth 40.699 1.06 0.11 0.01** 0.03** - 4 Firm Size 52.312 6.93 2.15 - 0.04** - 0.05** - 0.05** - 5 Distance to bankruptcy 52.774 1.03 4.97 - 0.02** - 0.01* 0.04** 0.24** - 6 Slack Index 50.015 0.03 0.01 0.02** 0.06** 0.02** - 0.33** 0.02** - 7 Env. Dynamism 40.713 0.02 0.02 0.01 - 0.01 0.01* 0.03** - 0.04** 0.01* - 8 Performance above aspirations (historical) 48.975 0.04 0.22 0.02** 0.05** 0.01** - 0.14** - 0.12** 0.01* 0.01* - 9 Performance below aspirations (historical) 48.975 -0.05 0.27 -0.01** 0.01** 0.05** 0.13** 0.64** 0.01** - 0.02** 0.04** - 10 Performance above aspirations (social) 53.219 0.12 0.17 0.01 0.03** 0.00 - 0.01** 0.08** 0.08** - 0.02** 0.46** 0.10** - 11 Performance below aspirations (social) 53.219 -0.06 0.30 - 0.01** 0.00 0.02** 0.20** 0.72** 0.04** - 0.01** -0.00 0.92** 0.12**

Table 2 Fixed effects panel regressions for high risk

Historical Aspirations Social Aspirations

Model 1 Model 2 Model 3 Model 4 Model 5

Variables Control SE Main SE Interaction SE Main SE Interaction SE

Industry R&D growth 0.522 0.294 0.633* 0.318 0.559** 0.319 0.564 0.295 0.546 0.295 Firm Size -0.839** 0.067 -0.841** 0.074 -0.842** 0.074 -0.834** 0.067 -0.838** 0.067 Distance to Bankruptcy -0.015 .021 -0.023 0.026 -0.023 0.026 -0.035 0.026 -0.036 0.026 Slack 90.547** 11.919 98.085** 14.164 96.909** 14.174 91.445** 11.953 90.616** 11.961 Dynamism -2.534 2.534 -2.667 2.537 -2.096 2.355 -2.435 2.362 Performance below aspirations -0.102 0.235 0.135 0.245 0.241 0.233 0.364 0.244 Performance above aspirations 0.383* 0.168 0.035* 0.167 0.202 0.196 0.193 0.196 Interaction 28.093* 0.319 21.415 11.671 Model F 63.98** 30.51** 27.26** 36.91** 32.72** R² .008 .009 .008 .007 .008 N 34.630 31.345 31.345 34.563 34.563

**. Correlation is significant at the 0.01 level *. Correlation is significant at the 0.05 level

Table 3 Fixed effects panel regression for low risk

**. Correlation is significant at the 0.01 level *. Correlation is significant at the 0.05 level

Historical Aspirations Social Aspirations

Model 1 Model 2 Model 3 Model 4 Model 5

Variables Control SE Main SE Interaction SE Main SE Interaction SE

Industry R&D growth 0.362** 0.097 0.373** 0.012 0.338** 0.104 0.399** 0.097 0.372** 0.097 Firm Size -0.178** 0.021 -0.146** 0.025 -0.146** 0.024 -0.176** 0.021 -0.175** 0.021 Distance to Bankruptcy 0.033** 0.007 0.015 0.009 0.015 0.009 0.019* 0.009 0.022 0.009 Slack 20.063** 3.972 8.453 4.741 8.06 4.739 20.401** 3.984 20.436** 3.986 Dynamism 0.016 0.834 -0.142 0.834 -0.106 0.771 -0.189 0.774 Performance below aspirations 0.284** 0.077 0.356** 0.079 0.152* 0.075 0.184* 0.079 Performance above aspirations 0.099 0.055 0.091 0.054 0.246** 0.066 0.245** 0.065 Interaction 12.796** 4.429 2.970 3.868 Model F 37.83** 14.21** 13.48** 24.33** 21.36** R² 0.005 0.004 0.004 0.006 0.006 N 34.124 30.951 30.951 34.060 34.060

Regression Analysis

Table 2 and 3 show the fixed effects panel regression results, once for high risk as dependent variable (table 2) and once for low risk as dependent variable (table 3). All regressions were done twice, once for performance compared to the historical aspirations of the organizations (model 2 and 3) and once for performance compared to the social aspirations of the organizations (model 4 and 5). The first pair of columns shows the results of the regression with the control variables. In model 2 and 4, the effects of the independent variables, negative performance feedback below historical and social aspirations, were added and model 3 and 5 explore the moderating effects of environmental dynamism.

The regression coefficients for performance below aspirations (historical and social) have to be interpreted carefully. Following the interpretation of previous studies that included performance below aspirations, a positive coefficient for performance below aspirations indicates that the further the performance drops below the aspiration level, the lower the engagement in low/high risk is. A negative coefficient on the other hand means that the further past performance falls below their aspirations, the higher the engagement is in low/high risk activities (Chen & Miller, 2007, Iyer & Miller, 2008, Lim & McCann, 2014).

Hypotheses 1a and b represented the baseline model, stating that a decrease in performance below aspirations would lead to less high risk (Hypotheses 1a) and more low risk (Hypothesis 1b). Thus a positive, significant coefficient is expected for the effect of negative performance feedback on high risk and a negative, significant coefficient of performance feedback for low risk. However, the coefficient of negative performance feedback in table 2 model 2 show no significant effect on high risk for aspirations below historical (t = - 0.04, p = 0.9) and social aspirations (t = 1.03, p = 0.3). Hypothesis H1a is thus not supported. Regarding the effect of negative performance feedback on low risk, table 3 model 2 shows that the coefficients for negative performance feedback are significant, but positive for both historical aspirations (t = 3.72, p < .01) and social aspirations (t = 2.01, p <

.01). This shows, that firms engage in significantly less low-risk activities, as performance falls further below their historical aspirations and social aspirations. This is contrary to what was expected in the baseline model, which hypothesized that a decrease in performance below the aspiration point (historical and social) leads to an increase in low risk. Hypothesis 1b is therefore also not supported.

Next, the interaction of performance below aspirations and environmental dynamism was added in order to explore the moderation effect of environmental dynamism. Hypothesis 2a states that environmental dynamism negatively moderates the relation between negative performance feedback and high risk, expecting the negative effect of performance feedback on high-risk activities to strengthen. Model 3 in table 2 shows, that the interaction between negative performance feedback based on historical aspirations and environmental dynamism is positive and significant (t = 2.11, p <.05). The interaction between the effect of negative performance feedback based on social aspirations and environmental dynamism however is not significant (p = .067). Hypothesis 2a is thus partly supported. A simple slope analyses was done in order to plot the interaction. To plot the interaction for firms performing below historical aspirations, negative performance feedback and environmental dynamism took the value of one standard deviation below (i.e. low level) and above (i.e. high level) the mean. The plot of the interaction is shown in figure 1. To illustrate the effect of negative performance feedback, -1 and 0 was used to present the values of performance below aspirations. The figure thus has to be read from right to left: when performance below aspirations drops from 0 to -1 it means that firms are experiencing further performance shortfalls relative to their aspirational level. Consistent with hypothesis 2a, figure 1 shows a negative relationship between negative performance feedback and high risk under highly environmental conditions (t = 1.29, p > .05). Moreover, it shows a positive relationship between negative performance feedback and high risk under when environmental dynamism

is low, meaning that organizations increase their high-risk activities as a response to negative performance feedback when the environment is more stable (t = -4.97, p >.01).

Figure 1 Moderating effect of environmental dynamism on the relationship between negative performance feedback based on historical aspirations and high risk.

Subsequently, the interaction of performance below aspirations and environmental dynamism was added in order to answer hypothesis 2b which states that environmental dynamism negatively moderates the relation between negative performance feedback and low risk in the sense that environmental dynamism turns the positive effect of negative performance feedback on low risk negative. However, the expected positive relationship of negative performance feedback and low risk of hypothesis 1b was not found (the effect was negative). The interaction of negative performance feedback based on historical aspirations and environmental dynamism is significant and positive (t = 2.98, p<.05). For social aspirations however, the effect is not significant (p = 0.23). This result partly supports hypothesis 2b, meaning that the negative relationship between negative performance feedback and low risk is strengthened only for firms experiencing negative performance feedback based on historical aspirations. To plot this interaction, the same is done as for the moderation effect on high risk: negative performance feedback and environmental dynamism took the value of

-1 0

Performance < Historical Aspirations

one standard deviation below (i.e. low level) and above (i.e. high level) the mean. The plot of the interaction is shown in figure 2. To illustrate the effect of negative performance feedback, -1 and 0 was used to present the values of performance below aspirations in figure 2. The figure thus has to be read from right to left: when performance below aspirations drops from 0 to -1 it means that firms are experiencing further performance shortfalls relative to their aspirational level. Consistent with hypothesis 2b, figure 2 shows a negative relationship between negative performance feedback and low risk under highly dynamic environmental conditions (t = 3.91, p<.01). Moreover, it shows that organizations don’t change their low-risk activities as a response to negative performance feedback when environmental dynamism is low (p = 0.23).

Figure 2 Moderating effect of environmental dynamism on the relationship between negative performance feedback based on historical aspirations and low risk.

Discussion

This research was conducted to investigate whether environmental conditions moderate the relationship between negative performance feedback and organizational risk-taking. Furthermore, this study distinguished between different levels of risk (low/high risk)

Performance < Historical Aspirations

-1 0

in order to get a more nuanced understanding of organizational risk-taking. The results of the study overall support the idea that organizational risk-taking after negative performance feedback below aspirations differs under different environmental conditions. In the following, I will explain and discuss the major findings, embed it in the existing literature and outline its contributions. Next, I will assess the limitations of this study and give recommendations for future research. Finally, I will end this paper by drawing a conclusion.

Major Findings

As outlined in the research gap, there is an ongoing discussion about the effect of negative performance feedback on organizational risk-taking. Specifically, there are two lines of argumentation. One group of researchers argues that organizations become more risk-seeking when their performance drops further below their aspirations. The underlying argument is that the need for improvement becomes more salient and organizations actively search for alternative courses of actions and accept a higher level of risk in order to achieve outcomes that meet their aspirations (Audia & Greve, 2006; Boyle & Shapira, 2012; Desai, 2008; Miller & Chen, 2004). The other stream of research argues for an effect in the opposite direction, namely that organizations respond more risk averse to performance shortfalls below aspirations as those are perceived as threats and organizations strive to prevent further organizational decline (Staw et al., 1981; Stikin & Pablo, 1992; Wiseman & Bromiley, 1996; McNamara & Bromiley, 1997, Greve, 2003; Audia & Greve, 2006; Iyer & Miller, 2008). Both studies have found empirical support which is why additional knowledge is needed concerning boundary conditions. This study attempts investigated whether environmental dynamism influences the organization’s perception of risk and specified conditions under which organizations become either more risk seeking or more risk averse.

The findings of this study provide rare evidence in support of the hypothesis of risk aversion when performance falls further below aspirations. The results suggest, that

organizations engage in less risk as performance drops further below their aspirations. This is contrary to what was expected in the baseline model and contrary to the predictions of the behavioral theory of the firm (Cyert & March, 1963) and prospect theory (Kahneman & Tversky, 1979). However, it is in line with previous studies (Audia & Greve, 2006; Sitkin & Pablo, 1992) offering support for the threat-rigidity literature which argues that performance shortfalls are perceived as a threat, creating a tendency in organizations to prefer well-learned routines in order to keep control over future performance (Staw et al., 1981). Yet, this effect was found only for low risk. The effect for high risk is negative indicating a lower engagement in high-risk activities as a response to negative performance feedback, however it was not significant. Greve (1998) also found that the effect of risk-taking as a response to negative performance feedback disappears for the riskiest type of change (Greve, 1998). A possible explanation for this is that organizations engage in high risk less frequently which makes it more difficult to find effects. However, the fact that different results were found for low and high risk supports the notion that risk-taking differs, depending on the level of risk concerning a particular activity. The general direction of this finding is in line with the argument made earlier based on the shifting-attention model of March and Shapira (1992) suggesting that organizations focus on the aspiration point in the face of low-risk decisions. If a low-risk investment fails, it might decrease organizational performance but it won’t threaten organizational survival, leading organizations to accept more risks. Performance shortfalls are seen as repairable gaps, which can be overcome with low-risk investments. When facing high-risk decisions on the other hand, organizations shift their attention to the survival point, which induces threat in managers because a failure of a high-risk investment might bring the organization close to bankruptcy.

Furthermore, it was expected in hypothesis 2a that the effect of negative performance feedback below aspirations on risk-taking is negatively moderated by environmental dynamism. The results are in line with this hypothesis, indicating that organizations refrain

from high risk as a response to declining organizational performance below when environmental dynamism is high. This result shows that environmental dynamism affects the perception of risk. It supports the proposed arguments based on the shifting-attention model that environmental dynamism intensifies the perception of threat and danger connected to performance shortfalls. Environmental dynamism is characterized by a high degree of uncertainty, which makes it hard to foresee the effects of future decisions, increasing the variability of possible outcomes and thus increasing the perception of risk in decisions (Eisenhardt & Bourgeois, 1988). Such conditions increase the difficulty to process information and trigger stress and anxiety in managers, leading them to prefer well-known routines in order to offset the uncertainty of the environment (Tushman, 1979). The chance for failure is bigger under highly dynamic environmental conditions, which is why managers focus shifts to the survival point in times of uncertainty. Subsequently, when the environment is more stable, their focus of attention shifts back to the aspiration point because organizations feel less threatened whenever the environment is characterized by a low degree of uncertainty. This in turn leads managers to accept higher levels of risk in order to conquer performance shortfalls.

Hypothesis 2b states that environmental dynamism negatively moderates the enhancing effect of negative performance feedback on engagement in low-risk choices in the sense that the effect turns negative whenever the environment is highly dynamic. The results show, that that low-risk engagements decrease further in response to performance shortfalls below aspirations when the environmental dynamism is high. Since this study did not find a positive effect of negative performance feedback on low risk in the baseline model, the findings support this hypothesis. This indicates that environmental dynamism induces threat perceptions in the face of high-risk and low-risk activities for organizations performing below aspirations in such a compelling way, that organizations refrain from all kinds of risk, even from activities that are considered to contain a lower degree of risk. This finding offers further

support for the argument that a highly dynamic environment shifts the focus of organizations to the survival point of the firms, which leads organizations to perceive risks as threats to organizational survival.

However, all of these effects were found solely for performance shortfalls below an organizations historical aspirations, not as a response to performance shortfalls below an organizations social aspirations. The fact that the effect was not found for social aspirations could indicate, that organizations rely stronger on historical aspirations. Whether organizations generally rely more on performance feedback based on historical aspirations or on performance feedback based on social aspirations is still in debate (Baum & Dahlin, 2007). According to Greve (2003), it depends on whether the organization feels like they are comparable to their reference group. If they are, they prefer social aspirations. If, however they evaluate themselves as being unique, they prefer historical aspirations (Greve, 2003). The construct of performance feedback based on social aspirations also holds some difficulties. We do not know exactly how organizations choose their reference group and whether they evaluate themselves as being similar or unique (Greve, 1998). Furthermore, it is possible that the functional background of firms is different which results in unexpected reference groups (Schurr, 1987). Following prior studies (Audia & Greve, 2006; Chen & Miller, 2004; Desai, 2008), this research took the average industry performance as a basis for the performance feedback measure based on social aspirations. However, it is possible that this does not reflect the reference group of organization in real life.

To summarize, the major finding of this study is, that organizations respond differently to a decline in performance below aspirations under different environmental conditions, which indicates that environmental dynamism affects the risk-perception of managers. A highly dynamic environment intensifies perceptions of danger, which is why performance shortfalls below aspirations lead organizations to become more risk averse. Under highly dynamic conditions, managers seem to prefer safer options, in order to offset the