Performance feedback and organizational behavior : the influence of differences in the dimensionality of performance measures

(1)

Final Thesis

PERFORMANCE FEEDBACK AND ORGANIZATIONAL BEHAVIOR:

THE INFLUENCE OF DIFFERENCES IN THE DIMENSIONALITY OF

PERFORMANCE MEASURES

ANOUK LISANNE VAN DER MEULEN University of Amsterdam

MSc. in Business Administration – Strategy Track Student number: 10868518

Supervisor: Silveira Barbosa Correia Lima, B. Date: 24/06/2016

(2)

STATEMENT OF ORIGINALITY

This document is written by Student Anouk Lisanne van der Meulen 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)

TABLE OF CONTENTS

Abstract ... 4

I Introduction ... 5

II Theoretical framework ... 10

1. Performance feedback theory ... 10

2. Attention-based view ... 16

3. Dimensionality of performance measures ... 19

III Method ... 23 1. Data ... 23 2. Dependent variable ... 23 3. Independent variables ... 24 4. Control variables ... 25 5. Model specification ... 26 IV Results ... 28

V Discussion and Conclusion ... 38

References ... 43

(4)

ABSTRACT

In response to the conflicting results that are found in studies following the Behavioral

Theory of the Firm on the relationship between performance feedback and organizational

behavior, this study proposes and tests differences in the dimensionality of performance measures as a new theoretical explanation for these contradictory findings. Building on the Attention-Based View of the Firm, this study argues that research in the field of performance feedback theory has overlooked the fact that organizations may respond to performance feedback differently, due to differences in the distribution of managerial attention. By introducing the notion of differences in the dimensionality of performance measures, this study aims to show that how firms respond to performance feedback depends on the performance dimension the firm has focused on in the determination of its performance targets. This study contributes to the existing literature by shifting the focus of performance feedback literature towards the mechanisms that are underlying organizational behavior. It provides a deeper understanding of the implications that may arise from using different measures for the same theoretical construct. Testing the theoretical predictions, a quantitative analysis is conducted on a sample of U.S. manufacturing firms from 1980 to 2010. The findings support to a degree the contention that some of the conflicting results that are found in performance feedback literature can be explained by the dimensionality of different performance measures. However, the findings also suggest that further research is needed on possible factors that influence the direction of organizational responses to performance feedback.

Key words: Behavioral theory of the firm, performance feedback, attention-based view, performance measures, dimensionality, quantitative analysis

(5)

I. INTRODUCTION

A central issue in organizational literature is how feedback on organizational performance affects strategic decision-making and organizational behavior (Cyert & March, 1963; Greve 2003ab, 2011; Audia & Greve, 2006; Iyer & Miller, 2008; Kim, Finklestein & Haleblian, 2015). The Behavioral Theory of the Firm (BTOF) by Cyert & March (1963), forms the underlying model for performance feedback theory. The BTOF states that organizational decision-making is influenced by the current performance of the firm relative to their predetermined aspiration levels. Aspiration levels serve as the firm’s target for performance, and can be based on past performance of the firm itself (historical aspirations) and the performance of its competitors (social aspirations) (Greve, 2003a). Decision makers can be defined as individuals who make decisions on behalf of the organization. In this process of decision-making they are constrained in their cognitive capabilities, for instance by bounded rationality and limited managerial attention (Ocasio, 1997; Greve, 2003a). The consequences of the strategic decisions they make based on the feedback on performance, are crucial to the decision makers since their careers depend on the outcome of their decisions (Greve, 2011).

Research has investigated the effect of performance feedback on different types of organizational behavior, such as organizational change, risk taking, capital expenditures, innovation, R&D expenditures, and mergers and acquisitions (Bromiley 1991; Greve, 2003b; Audia & Greve, 2006; Iyer & Miller 2008; Joseph & Gaba, 2015). According to performance feedback theory, performance above the aspiration level is interpreted as the applied strategy being successful. The firm therefore has little incentive to change its current behavior or to engage in risk taking decision-making. On the other hand, the effects of performance feedback below the aspiration level are the subject of an ongoing debate in the literature (Audia & Greve, 2006). Some studies, following the BTOF, find an increase in organizational change as a response to performance feedback below the aspiration level (Cyert & March, 1963; Bromiley

(6)

1991; Greve, 2003b, 2011), whilst others find a decrease in organizational change as a response (Audia & Greve, 2006; Iyer & Miller, 2008; Joseph & Gaba, 2015). The conflicting results for the effect of performance below the aspiration level on the direction of organizational change suggest that organizations may respond differently to performance feedback, which has led to a widespread of research in the field.

In order to investigate what causes firms to respond differently to performance feedback below the aspiration level, studies bring forward several moderators that may affect the relationship between performance feedback and organizational change (Greve, 2011). Analyzing the results of previous studies on performance feedback, some studies include the notion of differences in the focus of managerial attention that may cause the differences in the direction of organizational change in response to performance feedback (Audia & Greve, 2006; Iyer & Miller, 2008). Also Greve (2011) accurately points out that, although the effects of several moderators have been studied extensively, these studies overlook an important factor in their analysis. In his research, Greve (2011) argues that firms interpret strategic actions differently. What is interpreted by the one firm as an action of high risk, may be interpreted as less risky by the other firm, depending on a variety of contextual factors (Greve, 2011; Ocasio 1997). Therefore, studies should also take into consideration the content that is underlying strategic actions (Greve, 2011). Kacperczyk et al. (2015) argue that, first of all, a clear distinction has to be made between risk taking and organizational change. It is important to understand that risk taking and organizational change are two distinct concepts that can be interpreted differently by different organizations (Palmer & Wiseman, 1999; Kacperczyk et al., 2015). In The Behavioral Theory of the Firm, Cyert & March (1963) give no reference to this distinction, causing following studies to consequently assume that an increase in organizational change is always associated with an increase in taking risk (Kacperczyk et al., 2015). However, these studies should take into account that firms can take actions that are risky while at the same

(7)

time do not change the current strategy of the firm, and vice versa (Ocasio 1995, as cited in Greve, 2011).

Bromiley, Miller & Rau (2001) note that the BTOF only predicts whether or not a firm will change its current strategic behavior in reaction to performance feedback, however it does not specify the direction of change and its corresponding outcomes. In order to investigate the direction of change in performance feedback theory, Bromiley et al. (2001) emphasize that it is essential to understand the mechanisms underlying the strategic actions. Bromiley & Harris (2014) point out that the effects of performance feedback should be examined more theoretically-grounded, incorporating the implications that different measures of performance feedback may have on the direction of change. Performance feedback is defined as the evaluation of the firm’s current performance relative to a set of predetermined aspiration levels (Cyert & March, 1963). However, these aspiration levels are formulated reflecting a specific measurement of organizational performance (Bromiley & Harris, 2014).

In their research, Combs, Crook & Shook (2005) identified three dimensions of organizational performance, i.e. accounting measures, growth measures and stock market measures. Each performance measure consists of distinct boundaries and dimensionality, and consequently may have different implications for the relationship between performance feedback and organizational behavior (Combs et al., 2005). The notion of organizational performance consisting of different dimensions that each have their own specific characteristics, provides a renewing point of view on the existing literature studying the effects of performance feedback. When a firm focuses on a specific dimension of performance, their targets for performance are formulated based on the characteristics that correspond with that specific performance dimension. Different firms focusing on different dimensions of performance may result in differences between the performance targets each firm uses in the evaluation of their performance, which may cause differences in managerial decision-making.

(8)

Ocasio’s Attention-Based View of the Firm (1997) suggests that organizational behavior is determined by the distribution of the attention of the firm’s decision makers. That is, the strategic decisions they make depend on their focus of attention, which is based on the specific characteristics and structure of the firm and of its environment (Ocasio, 1997). Different measures of performance reflect different shareholder perspectives and, therefore, lead to different managerial attention (Bromiley et al., 2001). Accounting measures for example put focus on cost reduction, whereas growth measures put focus on expanding business. In this way, different performance measures shift managerial attention towards different problems, which in turn may lead to different directions of organizational change.

In response to these findings, this study proposes differences in the dimensionality of performance measures as a new theoretical explanation for the conflicting results that are found in literature examining the relationship between performance feedback and organizational behavior (Iyer & Miller, 2008; Greve 2011; Bromiley & Harris 2014; Kacperczyk et al., 2015). The research question of this study, therefore, is formulated as:

‘How do differences in the dimensionality of performance measures explain the

conflicting results that are found in studies on the relationship between performance feedback and acquisition behavior?’

This study focuses on acquisition behavior as a response to performance feedback, with return on assets (ROA), Tobin’s Q, growth, and market share being used as measures of the different dimensions of organizational performance. The research contributes to the existing literature in the field of organizational theory in the following ways: By introducing differences in the dimensionality of performance measures that are used in organizational theory, this study shifts the focus of performance feedback literature towards the mechanisms that are underlying

(9)

organizational behavior. It provides a deeper understanding of the implications that may arise from using different measures for the same theoretical construct (Combs et al., 2005; Bromiley & Harris, 2014). By means of a quantitative analysis on the effects of different dimensions of performance measures on the relationship between performance feedback and acquisition behavior, this study responds to a call from recent research (Audia & Greve, 2006; Iyer & Miller, 2008) to find an explanation for some of the variance that is found across findings in the field of performance feedback literature. A new insight on performance feedback theory is provided by combining the theoretical perspectives originating from the Behavioral Theory of the Firm by Cyert & March (1963) and the Attention-Based View of the Firm by Ocasio (1997) and expanding these two views through incorporating the notion of differences in the dimensionality of performance measures building on Combs et al. (2005). This study emphasizes the importance of clarity on the conceptual definition of multi-dimensional constructs such as performance. It encourages researchers to take notice of the underlying mechanisms of organizational behavior, taking into account factors that may influence the direction of search and organizational change.

This study is structured as follows: In chapter 2, the theoretical perspectives underlying performance feedback theory and the attention-based view are described. The chapter also explains and specifies the different dimensions of organizational performance that possibly influence the relationship between performance feedback and organizational behavior. The method of data collection and data analysis is presented and substantiated in chapter 3. In chapter 4 the results of the data analysis are presented and described. At last, the study is concluded in chapter 5, with a discussion and a summary of the most important findings and some recommendations for further research.

(10)

II. THEORETICAL FRAMEWORK

A central issue in organizational literature is the effect of performance feedback on organizational behavior. However, conflicting results are found in studies examining this relationship. Research on the relationship between performance feedback and organizational behavior is guided by performance feedback theory, suggesting that strategic decision-making in a firm is influenced by the firm’s current performance relative to the firm’s prior experience (Cyert & March, 1963). Alternatively, the attention-based view of the firm indicates that organizational behavior is determined by the distribution of the attention of the firm’s decision makers (Ocasio, 1997). The direction of managerial attention is affected by shareholders’ perspective on performance. In an attempt to explain the conflicting results in performance feedback theory, I combine and expand these two theoretical perspectives, by including the notion of differences in the dimensionality of performance measures (Combs et al., 2005). In the following sections, the distinct theoretical perspectives are summarized and a specification of the different performance measures is provided.

1. Performance feedback theory

Performance feedback theory originates from a combination of insights drawn from the Behavioral Theory of the Firm (BTOF) by Cyert & March (1963) and is grounded in the Carnegie school of thought (Greve, 2003a; Haleblian, Kim & Rajagopalan, 2006). The Carnegie school of thought has extended organizational theory through the application of search, limitations of managerial attention, bounded rationality, and satisficing behavior, on the concept of strategic decision-making (Ocasio, 1997; Greve, 2003b). In this view, the notion of rational choice is eliminated, assuming decision makers to be bounded in their capability to act rational (Simon, 1947, as cited in Ocasio, 1997). Bounded rationality is described by March & Simon (1958, as cited in Greve, 2003b) as the acknowledgement that decision makers have

(11)

limited information, attention and processing ability, causing the decision maker to be unable to maximize the outcomes of their decisions. Instead of maximizing outcomes, decision makers resort to satisficing behavior, implying the decision makers to ‘set a goal that they try to meet and evaluate alternatives sequentially until one that satisfies the goal has been found’ (Greve, 2003a: 12). As the performance goal is fulfilled, there is no need to search for an alternative any further. This process of searching and evaluating alternatives until the one that fulfills the performance goal has been found, is seen as learning from performance feedback. By learning from performance feedback, decision makers attempt to deal with the restrictions of bounded rationality (Greve, 2003a; Jordan & Audia, 2012). Building on the BTOF, Levitt & March (1988) define organizational learning as the integration of the firm’s perception of past experience, of the firm itself or other firms in the industry, into the routines that guide the organization’s behavior. How firms interpret and integrate this past experience depends on the performance targets, i.e. aspiration levels (Levitt & March, 1988).

Performance feedback theory specifies that the outcomes of prior behaviors influence the future behaviors of the firm (Cyert & March, 1963). According to the theory, firms adjust their behavior based on the evaluation of the firm’s observed performance relative to a set of

predetermined aspiration levels. An aspiration level can be seen as a target for performance, and is defined by Schneider (1992: 1053, as cited by Greve, 2003a) as “the smallest outcome that would be deemed satisfactory by the decision maker”. Aspiration levels play an important role in the decision-making process, serving as the boundary between perceived success and failure when evaluating firm performance (Lant & Shapira, 2008). Based on the outcome of the performance assessment, decision makers decide whether to maintain or change their current strategy. In determining the aspiration level, organizations incorporate both historical performance and social performance. Historical performance is defined as past performance of the firm itself, and can be used as a forecast for future performance. Social performance is

(12)

defined as the performance of competing firms in the industry, and can be used as an indicator of the firm’s capabilities in comparison to its direct competitors (Greve, 2003b; Joseph & Gaba, 2015). The difference between the firm’s observed performance and its aspiration levels is called the ‘attainment discrepancy’, and determines whether or not the firm will change its current organizational behavior (Iyer & Miller, 2008).

In the evaluation of performance, different aspiration-level models can be used, each reflecting different theoretical assumptions (Joseph & Gaba, 2015). For example, some studies apply a separate aspiration-level model, using historical and social aspirations as two independent indicators in comparing observed performance against aspirations (Joseph & Gaba, 2015). In this model, historical aspirations are measured as a weighted moving average of the firm’s past performance, whereas current mean industry performance is used as a measure of

social aspirations (Bromiley & Harris, 2014). The separate aspiration-level model allows for different directions for performance relative to the firm’s own performance versus performance relative to the industry. On the other hand, a weighted-average model for aspiration-levels combines the separate measures for both historical and social aspirations into a joint, single measure of aspirations (Joseph & Gaba, 2015). The weighted-average model forms a more realistic representation of corporate practice, assuming that firms use only one set of goals to evaluate performance, although these goals do reflect various factors (Bromiley & Harris, 2014). Testing the suitability of both the weighted-average and the separate aspiration-level model, Bromiley & Harris (2014) find strong support for the separate model in favor over the weighted-average model of aspiration levels. Since in this study I focus on differences in the directions of performance feedback, in combination with the results generated by Bromiley & Harris (2014), I will use a separate aspiration-level model for this research.

Performance feedback theory makes a distinction between performance above the aspiration level and performance below the aspiration level. A commonly accepted perspective

(13)

of studies adopting performance feedback theory is that, as performance rises above the aspiration level, the firm will be less likely to change its current organizational behavior (Greve, 2003ab; Audia & Greve, 2006; Haleblian et al, 2006; Iyer & Miller, 2008; Greve, 2011). Performance feedback above the aspiration level is interpreted by the decision makers as a validation for the previously applied organizational strategy to be successful. They receive little incentive to change their current behavior and will be more likely to persist in exploiting the behavior that has proven to be successful. In other words, the firm becomes more resistant to change and will avoid taking risky decisions that could harm its current position. Based on this argument, the following hypothesis is proposed:

Hypothesis 1: Acquisition count decreases with firms' past performance above aspirations when focusing performance measured by either ROA, Tobin’s Q, growth, or market share.

Empirical research on the effects of performance feedback below the aspiration level however, have found conflicting results (Audia & Greve, 2006). Some studies find an increase in organizational change as a response to low performance (Cyert & March, 1963; Bromiley 1991; Greve, 2003b, 2011), whilst others find evidence that supports the notion of organizational change to decrease as a response to low performance (Audia & Greve, 2006; Iyer & Miller, 2008; Joseph & Gaba, 2015). As performance falls below the aspiration level, performance feedback theory predicts that firms will be incentivized to change its current organizational behavior. Intending to overcome the performance deficiencies, decision makers will search for more viable alternative opportunities and make changes in the current strategy (Cyert & March, 1963). Cyert & March (1963: 121) describe this search process as problemistic search, referring to it as “search that is stimulated by a problem […] and is directed toward finding a solution

(14)

to that problem”. Problemistic search is most likely conducted in areas near the perceived problem or in areas that have previously brought forward a solution to a similar problem (Argote & Greve, 2007). Problemistic search processes can lead to, inter alia, an increase in capital expenditures, innovation, expenses in R&D, or mergers and acquisitions (Bromiley 1991; Greve, 2003b; Audia & Greve, 2006; Iyer & Miller 2008; Joseph & Gaba, 2015). For instance, problemistic search may lead to an increase in R&D expenditures as a potential solution for the firm’s performance deficiencies through the exploration of ways to improve existing products or to create new products (Greve, 2003b).

Studying the effect of performance feedback on R&D expenditures in the shipbuilding industry, Greve (2003b) finds that that low performance leads to an increase in R&D expenditures in the shipbuilding industry. The study shows that, in case of performance below the aspiration level, firms tend to search for viable alternatives that will increase performance, resulting in an increase in innovation launches. However, on the moderating effect of organizational slack on this relationship, no evident results are found (Greve, 2003b). Also Haleblian et al. (2006) predict that performance below the aspiration level leads to problemistic search and risk taking behavior, trying to find alternative acquisition opportunities. The probability of the firm continuing its particular acquisition strategy therefore will decline. In their empirical analysis of the U.S. commercial banking industry, Haleblian et al. (2006) found support for their predictions. The results of both studies are in line with the theory of Cyert & March (1963), assuming that performance below the aspiration level causes decision makers to induce a process of problemistic search, in an attempt to find an alternative way to reposition the firm and to increase performance (Greve, 2011).

Arguing that differences in firm size may lead to different perceptions on performance below the aspiration level, Audia & Greve (2006) propose firm size as a moderating variable on the relationship between performance feedback and risk taking behavior. Studying the

(15)

effects of performance feedback on organizational risk taking behavior, they find evidence that rejects the claim of risk seeking in response to negative performance feedback. As performance falls below the aspiration level, the results indicate a decrease in the level of risk taking within small firms, while risk taking within large firms is not affected by low performance or in some cases even increases. In an attempt to explain these results, Audia & Greve (2006) state that small firms may perceive low performance as a threat to firm survival and will decrease risk taking, while large firms may perceive low performance as a repairable gap that can be overcome. Also contrasting the BTOF, Iyer & Miller (2008) find that performance below the aspiration level negatively affects acquisition behavior. Investigating acquisition behavior as a response to performance feedback within U.S. manufacturing firms, they find that the hazard of acquisitions decreases as performance falls below the aspiration level. Based on this, Iyer & Miller (2008) argue that low performance may direct managerial focus towards solutions that solve problems within the existing business, rather than towards changing the organizational strategy and engaging in an acquisition. Incorporating the moderating effect of organizational slack on the relationship between performance feedback and acquisition behavior, a positive effect for unabsorbed and potential slack on the relationship between performance below the aspiration level and acquisition behavior is found. For performance above the aspiration level, Iyer & Miller (2008) find that the hazard of acquisitions decreases as performance increases. Also this is contrary to what would be expected from the BTOF by Cyert & March (1963).

The conflicting results that are found by studies investigating the effects of performance feedback on organizational behavior suggest that organizations may respond differently to performance feedback. In order to find out what causes firms to respond differently to performance feedback, studies have proposed a variety of moderators that may affect the relationship between performance feedback and organizational change (Greve, 2011). However, it remains unclear what causes the conflicting directions of change as a reaction to

(16)

performance below the aspiration level. Both Audia & Greve (2006) as well as Iyer & Miller (2008) suggest differences in the focus of managerial attention as a potential cause for the differences in the direction of organizational change in response to performance feedback. In their research, for example, Iyer & Miller (2008) use return on assets (ROA) as a measure for performance. Since ROA indicates the profitability of the firm relative to its total assets, in case of performance below the aspiration level, decision makers that focus on ROA as a measure for performance may decide to cut costs and decrease investments in acquisitions in order to increase performance, i.e. increase ROA. The following section introduces the attention-based view by Ocasio (1997) and connects this theoretical perspective to the results that are found in the performance feedback literature.

2. Attention-based view

On the conflicting results in the performance feedback literature, Greve (2011) accurately points out that these studies have overlooked the fact that organizations may interpret strategic actions differently. What is interpreted by the one firm as a rigorous change that is accompanied with high risk, may be interpreted as being a standardized and less risky procedure by the other firm, depending on a variety of contextual factors (Greve, 2011; Ocasio 1997). For instance, the decision to invest more in R&D is commonly accepted as a form of organizational change. However, for a firm that maintains a strategy that is highly focused on innovations, the decision to invest less in R&D may also be a form of organizational change. Put differently, whether investing more in R&D is perceived by the firm as a form of organizational change depends on the context of the firm. This reasoning implies the important notion that organizational change can take place in different directions. Greve (2011), therefore, argues that studies should also take into consideration the content that is underlying strategic actions in the determination of organizational change.

(17)

Kacperczyk et al. (2015) argue that, first of all, a clear distinction has to be made between risk taking and organizational change. It is important to understand that risk taking and organizational change are two distinct concepts that can be interpreted differently by different organizations (Palmer & Wiseman, 1999; Kacperczyk et al., 2015). In the BTOF, Cyert & March (1963) address the processes of problemistic search and organizational change. In explaining these processes, no reference is given to the difference between change and risk, causing studies that build on this theory to inattentively have assumed that an increase in organizational change is always associated with an increase in taking risk (Bromiley, 1991; Kacperczyk et al., 2015). As stated by Ocasio (1995, as cited in Greve, 2011) however, firms can take actions that are risky but at the same time do not lead to a change in the current strategy of the firm, and vice versa. The actions a firm undertakes as a result from the process of problemistic search thus can have different directions that will not always lead to an increase in risk taking behavior (Palmer & Wiseman, 1999; Kacperczyk et al., 2015). In the determination of the actual organizational change that is induced in response to performance feedback, this distinction should be taken into account. Also Bromiley, Miller & Rau (2001) support the argument of Kacperczyk et al. (2015), stating that the BTOF is limited in terms of the determination of organizational behavior in response to performance feedback. They notice that the theory by Cyert & March (1963) particularly predicts whether or not a firm will change its current strategic behavior in reaction to performance feedback, however it does not specify the direction of change and its corresponding outcomes. In order to investigate the direction of change in performance feedback theory, the underlying mechanisms need to be understood (Bromiley et al., 2001).

In Towards an Attention-Based View of the Firm, Ocasio (1997) puts forward that organizational behavior depends on the distribution of the attention of their decision makers. Attention is defined by Ocasio (1997) as the channelization of the decision maker’s time and

(18)

effort towards both issues and answers. The selection of a strategic decision depends on the issues and answers the decision maker focuses on (Ocasio, 1997). As proposed by the Carnegie school of thought, organizations have to deal with limitations of managerial attention regarding the process of strategic decision-making (Greve, 2003a). Since decision makers do not have complete information on the available actions, the consequences of these actions, and the available alternatives, the strategic decisions they make depend on the specific issues or answers they focus their attention on (Ocasio, 1997). The distribution of managerial attention is based on the specific characteristics and structure of the firm and its environment. Different organizational characteristics and structures may lead to a different attentional focus of the decision maker. The differences in the distribution of managerial attention in turn result in differences in organizational decision-making and strategic action (Ocasio, 1997).

One important characteristic of a firm is the way organizational performance is measured and evaluated (Combs et al., 2005). How the firm defines organizational performance influences the decisions that are made based on the evaluation of this specific type of performance. Different performance measures put emphasis on different problems within the business. Managerial attention subsequently is directed towards that specific problem, influencing the direction of the problemistic search process. In this way, building on the attention-based view of the firm, how the firm defines organizational performance may cause organizations to respond differently to performance feedback. Bromiley & Harris (2014) point out that the effects of performance feedback, therefore, should be examined in a more theoretically-grounded way, incorporating the implications of the use of different measures for performance feedback on the direction of organizational change. The following section explains and specifies differences in the dimensionality of performance measures that may be of influence on the different directions of organizational behavior in response to performance feedback.

(19)

3. Dimensionality of performance measures

Performance feedback is defined as the evaluation of the firm’s current performance relative to a set of predetermined aspiration levels (Cyert & March, 1963). Performance is known to be a multi-dimensional construct that can be measured in a wide variety of ways (Bromiley et al., 2001; Combs et al., 2005). However, aspiration levels that are used to assess performance often reflect one single dimension of organizational performance, directing managerial attention towards a specific target (Bromiley et al., 2001; Bromiley & Harris, 2014). Combs et al. (2005) state that developing a clear definition of the boundaries, dimensionality, and appropriate measures for the construct of organizational performance, is of high importance for future strategic management research. Although organizational performance is seen as one of the most important constructs in organizational literature, little attention has been paid to the development of a common understanding regarding its conceptualization and measurement. Without any conceptual clarity on organizational performance, measurement problems will arise, leading to an inordinate variance across the empirical findings in the field of organizational literature (Combs et al., 2005)

In their research, Combs et al. (2005) have identified three dimensions of organizational performance, i.e. accounting measures, growth measures and stock market measures. Each performance measure consists of distinct boundaries and dimensionality, having distinct implications for the relationship between performance feedback and organizational behavior (Bromiley et al., 2001; Combs et al., 2005). The different dimensions of performance reflect different shareholder preferences, influencing the distribution of managerial attention. In this way, different dimensions of performance result in different targets for performance, which may lead managerial attention towards different strategic actions in order to obtain their specific performance target (Ocasio, 1997; Bromiley et al., 2001; Bromiley & Harris, 2014). Consequently, when evaluating current performance relative to the aspiration level, each

(20)

dimension of organizational performance may have different effects on the distribution of managerial attention in the process of decision-making (Bromiley et al., 2001, Bromiley & Harris, 2014). As a result, the one dimension of performance may result in a different direction of organizational change than the other.

In research on performance feedback theory, most studies rely on accounting measures as a dimension of organizational performance (Bromiley & Harris, 2014). Most commonly, studies use return on assets (ROA) as an accounting measure for performance, defined as income (e.g. net income or pre-tax income) divided by total assets. However, the use of accounting measures, such as ROA, as an indicator of performance has been criticized. Managerial incentives often are based on accounting measures of performance, and since this performance dimension highly depends on accounting choices of the firm that can easily be manipulated, some researchers argue for a more sophisticated measure of performance such as Tobin’s Q (Bromiley & Harris, 2014). Tobin’s Q is an example of stock market measures of organizational performance, and is defined as the market value of the firm relative to its replacement costs (Lang & Stulz, 1993; Bromiley & Harris, 2014). In research, Tobin’s Q is often mentioned as the market to book value (Combs et al., 2005). Another distinct dimension of performance is growth, most commonly measured as the firm’s growth in sales (Combs et al., 2005). At last, despite of it not being an actual measure of performance, due to its frequent use in studies on organizational performance, Combs et al. (2005) also include a measure of market share in their research. Market share can be defined as an indicator of the relative size of the firm within its industry (Combs et al., 2005).

Each measure of organizational performance may have different implications for the distribution of managerial attention and, therefore, for the direction of organizational change. Accounting measures, providing information on the financial health of the organization, direct managerial focus towards cost reducing and profit maximizing strategies (Combs et al., 2005).

(21)

For example, when focusing on accounting measures as a dimension of performance, managerial attention is led towards maximizing the financial health of the organization on the short term. In case of low performance, a firm that focuses on ROA as a measure for performance might decide to cut costs by, for instance, decreasing expenditures on acquisitions in order to increase profits relative to aspirations. Based on this reasoning, the following hypothesis is proposed:

Hypothesis 2: Acquisition count decreases with firms' past performance below aspirations when focusing on performance measured by ROA.

When focusing on growth measures as a dimension of performance, on the other hand, managerial attention is led to expanding the business (Combs et al., 2005). In this case, a firm might decide to sacrifice profits and invest in opportunities for expansion in order to increase firm growth relative to aspirations on the long term (Combs et al., 2005). This is also the case for firms that focus on market share as a dimension of performance. Market share as a measure of performance provides information on the firm’s sales revenue relative to the total amount of

sales revenue in the industry (Joseph & Gaba, 2015). When focusing on market share, managerial attention is led to increasing sales revenue. As performance falls below the aspiration level, the firm might decide to invest more in opportunities that may expand its share in the market. Firms that focus on stock market measures as a dimension for performance, e.g. Tobin’s Q, channelize managerial attention towards increasing the firm’s stock value. In case of performance below the aspiration level, firms that focus on Tobin’s Q as a measure of performance might decide to increase investments. For increasing either firm growth, market share or Tobin’s Q, a possible way to achieve this is by investing in acquisitions. Taken together, these arguments lead to the following hypothesis:

(22)

Hypothesis 3: Acquisition count increases with firms' past performance below aspirations when focusing on performance measured by Tobin’s Q, growth, or market share.

In investigating whether the different dimensions of performance provides an explanation for some of the conflicting results that are found in performance feedback literature, the research of Iyer & Miller (2008) on the relationship between performance feedback and acquisition behavior is used as a guideline for this study. Building on the BTOF, Iyer & Miller (2008) predict that in case of performance below the aspiration level, the firm is more likely to search for alternative opportunities, resulting in an increased probability of acquisition. Contrary to their theoretical predictions, Iyer & Miller (2008) find in their analysis that the probability of acquisition decreases along with performance below the aspiration level.

However, by using return on assets (ROA) as a measure of performance, the results of their analysis might be biased towards shortterm profit maximization and cost reduction. Based on the theoretical perspective that is presented in this chapter, in case of performance feedback below the aspiration level, a firm that focuses on ROA as a dimension of performance is stimulated to cut costs and decrease investments in acquisitions in order to increase performance relative to the aspiration level. This perspective would provide an explanation for the conflicting results that Iyer & Miller (2008) have found in their analysis. In order to test the theory that is proposed in this section, in the following chapters a quantitative research is conducted, studying the effects of differences in the dimensionality of performance measures on the relationship between performance feedback and acquisition behavior. In this research, the four performance measures that are discussed in this section, i.e. ROA, Tobin’s Q, growth, and market share, are tested, building on the results that are found by Combs et al. (2005).

(23)

III. METHOD 1. Data

The dataset that is used for this research is derived from the CRSP-Compustat merged database, providing annual financial data and operational information on all publicly traded U.S. companies. Consistent with previous research on performance feedback and acquisition behavior (Iyer & Miller, 2008; Chen & Miller, 2007), the analysis is restricted to U.S. manufacturing companies with SIC codes from 2000 to 3999, minimizing the side effects that arise from differences in industry background. Industries that contain less than 5 firms are eliminated, in order to avoid possible biases in estimating the social aspiration level. For the analysis, data is used from 1980 to 2010 in fiscal years. Missing values are excluded from the dataset, in order to minimize bias during the regression analysis. Since this study builds on the conflicting results that are found by Iyer & Miller (2008), data similar to the dataset that is used in the research of Iyer & Miller (2008) is collected.

2. Dependent variable Acquisition behavior

The dependent variable that is investigated in this study is acquisition behavior, for this research in particular specified as acquisition count. Acquisition count is measured as the frequency of acquisitions performed in one fiscal year. The frequency of acquisitions is a count variable; a random variable consisting of non-negative integer values referring to the number of times an event occurs in a given interval of time (Cameron & Trivedi, 2013). The acquisition count variable is skewed to the right, being asymmetrically distributed. The mean of acquisition count lies at 0.27 acquisitions per fiscal year, and the standard deviation is 0.80 (see table 2).

(24)

3. Independent variables Performance

Four distinct dimensions of performance are tested in this research: performance measured by return on assets (ROA), Tobin’s Q, growth, and market share. All measures are calculated for the period t – 1. Return on assets (ROA) is calculated by dividing pre-tax income by total assets, with total assets being larger than 1. In this research, pre-tax income is favored over net income when calculating ROA, in order to avoid accounting issues that arise when using this more conventional measure of net income (Bromiley & Harris, 2014). Tobin’s Q is measured as the firm’s market value divided by its replacement costs (Lang & Stulz, 1993). Growth is defined as the firm’s growth is sales (Lucas, Knoben & Meeus, 2015). At last, market share is calculated by the firm’s sales revenue divided by the total sales revenue in its industry (Joseph & Gaba,

2015). All measures will be used in separate analyses in order to investigate their independent effects on the relationship between performance feedback and acquisition behavior.

Aspirations

Cyert & March (1963) describe aspirations to be determined by both a firm’s own past performance and the performance of other firms in its industry. Ambiguous results are found on how firms weight historical performance relative to social performance in the determination of their aspiration level (Baum et al., 2005; Chen & Miller, 2007). Therefore this study will use a separate aspiration-level model, testing for both historical and social aspirations separately (Bromiley & Harris, 2014). Historical aspirations are measured as the firm’s past performance over the period t – 2. Social aspirations are measured as the median performance of firms in the same 3-digit SIC industry codes over the period t – 2 (Chen & Miller, 2007; Iyer & Miller, 2008). Performance feedback, subsequently, is calculated as the difference between the firm’s actual performance and the aspiration level, either social or historical. In order to be able to test

(25)

acquisition behavior as a response to performance feedback above as well as below aspirations, a separate variable for both performance feedback above the aspiration level and performance feedback below the aspiration level is generated, by splitting each performance variable into two separate variables (Greve, 2003b). Performance below aspirations is measured as the difference between performance and aspirations for each observation in which the firm’s performance is less than the aspiration level. For the observations in which the firm’s

performance is greater than the aspiration level, performance below aspirations equals 0. Performance above aspirations, on the other hand, is equal to the difference between performance and aspirations for the observations in which performance is greater than the aspiration level. In case of performance being less than the aspiration level, performance above aspirations equals 0 (Greve, 2003b; Chen & Miller, 2007; Gaba & Joseph, 2013).

4. Control variables

Based on prior research, firm size is included to the analysis as a control variable. In this study firm size is computed as the log of employees (Greve, 2003b; Gaba & Joseph, 2013). Following the research by Audia & Greve (2006), firm size is also tested for having any moderating effects on the effect of different performance dimensions on the relationship between performance feedback and acquisition behavior. For this, separate interaction terms are generated.

In order to control for resource levels, measures of potential slack and unabsorbed slack are also added to the analysis. Potential slack is calculated by the debt to equity ratio (Bromiley, 1991; Iyer & Miller, 2008, Greve, 2011). Unabsorbed slack is calculated by the current ratio, i.e. current assets divided by current liabilities (Iyer & Miller, 2008). Since acquisition decisions are made based on feedback of past performance, in this research, the independent variables and the control variables are lagged one year relative to the dependent variable, i.e. acquisition

(26)

count (Chen & Miller, 2007; Iyer & Miller, 2008; Greve, 2011). A full list of the study variables and the definitions that are used in this research is presented in table 1.

5. Model specification

In testing the formulated hypotheses, random as well as fixed effects panel regression models are performed. However, both specifications provide substantially similar results (see Appendix 1). Therefore, further analysis is performed using only fixed effects models, because this specification is used more commonly in prior literature on this specific subject. Since the analysis involves count data, both Poisson regression analyses and negative binomial regression analyses are appropriate (Gaba & Joseph, 2013; Joseph & Gaba, 2015). A Poisson regression analysis assumes the mean and variance to be the same. In case of the variance being greater than the mean, called over dispersion, a negative binomial regression analysis is more appropriate (Cameron et al., 2013). Since the variance (.7954441) of acquisition count is greater than the mean (.2716604), the negative binomial regression analysis seems to be more

Table 1

Definitions of the Study Variables

Variable Definition Key Reference

Dependent Variable

Acquisition count Frequency of acquisitions performed in 1 fiscal year Greve (2011)

Independent variables

ROA Pre-tax income divided by total assets Bromiley & Harris (2014) Tobin's Q Firm market value divided by replacement costs Lang & Stulz (1993)

Growth Sales growth Lucas et al. (2015)

Market share Sales revenue divided by total industry sales revenue Joseph et al. (2015) Historical aspirations Past performance over the period t – 2 Chen & Miller (2007) Social aspirations median industry performance over the period t – 2 Chen & Miller (2007)

Control variables

Firm size Log of employees Greve (2003b)

Potential slack Debt to equity ratio Bromiley & Miller (1991) Unabsorbed slack Current assets divided by current liabilities Iyer & Miller (2008)

(27)

appropriate in this specific research (see table 2). However, the value of the Pearson dispersion statistic being close to 1.0 (1/df Pearson = .6338) rejects a case of over dispersion, indicating that Poisson regression analysis fits best to be used in this research (Woolridge, 2010). Therefore, Poisson panel regression analyses will be applied in this study.

(28)

IV. RESULTS

In table 2, the mean, standard deviation and correlations of the study variables are presented. On average, manufacturing firms engaged in 0.27 acquisitions per fiscal year, varying significantly across firms. The correlations show that acquisition count is positively related to ROA (.0524, p < .01) and market share (.01252, p < .01), while negatively related to Tobin’s Q (-.0131, p < .01), and insignificantly related to growth. The positive relationship of acquisition count with ROA and the negative relationship of acquisition count with Tobin’s Q, are in line with hypotheses 2 and 3 that are formulated based on theory. However, the correlation coefficients of the respective variables are relatively small.

Table 2

Mean, Standard deviation and Correlations of Study Variables

In order to test how the use of different performance measures explain the conflicting results that are found in studies on the relationship between performance feedback and acquisition behavior, fixed effects Poisson panel regression models are used. Table 3 shows the results of the Poisson regression analyses for acquisition count and performance feedback (below and above aspirations) focusing on ROA as a dimension of performance. Table 4 shows the results for performance feedback focusing on Tobin’s Q as a performance dimension. The results for growth and market share as performance dimensions are shown in table 5 and table 6. The analyses are performed using both social aspirations as well as historical aspirations in separate analyses. Model 1 includes only the control variables. As can be seen from the results, both firm size and unabsorbed slack are positively related to acquisition count (p < .001). Potential

M SD 1 2 3 4 5 6 7 8 1. count .2716604 .7954441 1.0000 2. ROA -.0549336 .6322355 0.0524* 1.0000 3. Tobin's Q 2.241141 3.673248 -0.0131* -0.4063* 1.0000 4. growth .6286314 45.2618 -0.0002 -0.0014 0.0113* 1.0000 5. market share .0538963 .1404054 0.1252* 0.0831* -0.0691* -0.0047 1.0000 6. firm size 6.72971 2.1829 0.2401* 0.2583* -0.2084* -0.0136* 0.4465* 1.0000 7. potential slack .504689 21.7745 0.0021 0.0054 -0.0040 -0.0001 0.0063 0.0113* 1.0000 8. unabsorbed slack 3.65399 7.27694 -0.0454* -0.0101* 0.0520* 0.0009 -0.0824* -0.2261* -0.0043 1.0000 * Correlation is significant at the 0.01 level (2-tailed)

Table 2

(29)

slack, however, was insignificantly related to acquisition count. Model 2 tests for the relationship between acquisition count and performance feedback measured by social aspirations. Model 3 tests for the relationship between acquisition count and performance feedback measured by historical aspirations. After having tested for model 1, 2 and 3, based on the results that are generated from the analyses, a fourth model is included to the research. In order to control for any moderating effects of firm size on the effects of different performance dimensions on the relationship between performance feedback and acquisition behavior, the interaction of firm size and performance feedback is included in model 4 (for ROA and Tobin’s Q). For this, two distinct interaction terms are generated, i.e. performance below aspirations multiplied by firm size, and performance above aspirations multiplied by firm size. Following the research by Audia & Greve (2006), model 4 tests whether the effect of different performance dimensions on the direction of change in response to performance feedback is different for smaller or larger firms.

In table 3 the results of the regression analysis for acquisition count and performance feedback focusing on ROA as a dimension of performance are shown. The results show that performance feedback focusing on ROA as a performance dimension is significantly related to acquisition count (p < .001) for both social aspirations (model 2) and historical aspirations (model 3). In order to investigate the effects different dimensions of performance on the direction of change as a response to performance feedback, the results of the regression analyses are divided in performance feedback above and below the aspiration level. As can be seen in table 3, for performance above the aspiration level, the results for model 2 (.616, p < .001) as well as for model 3 (.246, p < .001) are positive and significant. This indicates that, as performance rises above the aspiration level, acquisition count increases. These results contradict hypothesis 1, stating that acquisition count decreases with performance feedback

(30)

above aspirations. As performance falls below the aspiration level, table 3 shows that acquisition count decreases for model 2 (.592, p < .001) as well as for model 3 (.980, p < .001).

Table 3

Poisson Regression Model for Acquisition Count (performance = ROA)

Model 1 Model 2 Model 3 Model 4

Performance below aspirations

(ROA - social) 0.592*** -0.407***

(-0.47) (-5.62)

Performance above aspirations

(ROA - social) 0.616*** 0.832***

(6.36) (3.69)

Performance below aspirations

(ROA - historical) 0.980***

(9.80) Performance above aspirations

(ROA - historical) 0.246***

(3.54) Performance below aspirations

(ROA - social) x firm size 0.247***

(ROA - social) x firm size -0.0301

(-1.17) Firm size 0.205*** 0.203*** 0.208*** 0.218*** (16.39) (16.06) (15.26) (16.53) Potential slack -0.000286 -0.000268 -0.000441 -0.000273 (-0.50) (-0.47) (-0.69) (-0.48) Unabsorbed slack 0.0179*** 0.0176*** 0.0239*** 0.0171*** (6.80) (6.36) (6.89) (6.17) Number of observations 43274 42969 39317 42969 Number of firms 3038 3030 2785 3030

t statistics in parentheses. All independent variables are lagged by one year. * p<0.05, ** p<0.01, *** p<0.001

An illustration of the results of the regression analysis for performance feedback focusing on ROA as a dimension of performance and acquisition count is provided in figure 1. As can be seen in figure 1, the positive coefficients for acquisition count and performance below aspirations indicate that acquisition count decreases with performance below the aspiration level. In other words, the more performance decreases below the aspiration level, the less a firm

(31)

engages in acquisitions. These results support hypothesis 2, stating that acquisition count decreases as performance falls below the aspiration level, when focusing performance measured by ROA.

Figure 1

Regression Results for Performance Feedback (ROA) and Acquisition Count

The results of the regression analysis for acquisition count and performance feedback using Tobin’s Q as a dimension of performance are presented in table 4. Also performance feedback focusing on Tobin’s Q as a performance dimension is significantly related to acquisition count (p < .001). For this performance dimension, the results of both model 2 (.0272, p < .001) and model 3 (.0321, p < .001) are positive and significant in case of performance above the aspiration level. Similar to the results of performance feedback using ROA as a performance dimension and contrasting hypothesis 1, these results suggest an increase in acquisition count as a response to performance above the aspiration level. As performance falls below the aspiration level, table 4 shows that the results for model 2 (.107, p < .001) as well as for model 3 (.0318, p < .001) are positive and significant.

below aspirations 0 above aspirations

a cq u is it io n c o u n t

obtained results (ROA)

hypothesized results acquis it ion c oun t

(32)

Table 4

Poisson Regression Model for Acquisition Count (performance = Tobin's Q)

Model 1 Model 2 Model 3 Model 4

Performance feedback below

aspirations (Tobin's Q - social) 0.107*** 0.0469

(8.00) (0.98)

Performance feedback above

aspirations (Tobin's Q - social) 0.0272*** -0.0426**

(6.66) (-2.89)

aspirations (Tobin's Q - historical) 0.0318*** (3.82) Performance feedback above

aspirations (Tobin's Q - historical) 0.0321*** (5.93) Performance below aspirations

(Tobin's Q - social) x firm size 0.00755

(Tobin's Q - social) x firm size 0.0120***

(5.38) Firm size 0.205*** 0.224*** 0.218*** 0.219*** (16.39) (17.64) (15.96) (16.63) Potential slack -0.000286 -0.000299 -0.000411 -0.000305 (-0.50) (-0.53) (-0.67) (-0.54) Unabsorbed slack 0.0179*** 0.0201*** 0.0269*** 0.0195*** (6.80) (7.45) (7.85) (7.23) Number of observations 43274 42965 39274 42965 Number of firms 3038 3032 2785 3032

The results of the analysis focusing on Tobin’s Q as a dimension of performance are illustrated in figure 2. As can be seen in figure 2, the positive coefficients for acquisition count and performance below aspirations point to a decrease in acquisition count as performance feedback falls below the aspiration level. These results are against hypothesis 3, which states that, when using performance measured by Tobin’s Q, acquisition count increases as performance falls below the aspiration level.

(33)

Figure 2

Regression results for performance feedback (Tobin’s Q) and acquisition behavior

The results for acquisition count and performance feedback using growth and market share as a dimension of performance are presented respectively in table 5 and table 6. For both performance dimensions, the results that are generated from the regression analyses are not significant. These results suggest that acquisition count is not affected by performance feedback which is measured focusing on either growth or market share as a dimension for performance. Therefore, both analyses reject hypothesis 1 and hypothesis 3, stating that, when using performance measured by growth or market share, acquisition count decreases as performance rises above the aspiration level and increases as performance falls below the aspiration level.

a cqui si ti o n co unt

obtained results (Tobin's Q)

(34)

Table 5

Poisson Regression Model for Acquisition Count (performance = growth)

Model 1 Model 2 Model 3

aspirations (Growth - social) 0.00248 (0.69) Performance feedback above

aspirations (Growth - social) 0.0000823 (0.36) Performance feedback below

aspirations (Growth - historical) 0.000604 (0.27) Performance feedback above

aspirations (Growth - historical) 0.0000551 (0.22) Firm size 0.205*** 0.206*** 0.202*** (16.39) (15.01) (13.80) Potential slack -0.000286 -0.000345 -0.000459 (-0.50) (-0.58) (-0.73) Unabsorbed slack 0.0179*** 0.0282*** 0.0347*** (6.80) (6.51) (7.53) Number of observations 43274 38681 35324 Number of firms 3038 2741 2516

Table 6

Poisson Regression Model for Acquisition Count (performance = market share)

Model 1 Model 2 Model 3

aspirations (Market share - social) 0.487 (1.81) Performance feedback above

aspirations (Market share - social) 0.148 (1.18) Performance feedback below

aspirations (Market share - historical) 0.905** (2.79) Performance feedback above

aspirations (Market share - historical) 0.431 (1.60) Firm size 0.205*** 0.200*** 0.201*** (16.39) (15.47) (14.84) Potential slack -0.000286 -0.000290 -0.000403 (-0.50) (-0.51) (-0.65) Unabsorbed slack 0.0179*** 0.0186*** 0.0225*** (6.80) (6.91) (6.68) Number of observations 43274 43062 39393 Number of firms 3038 3034 2789

(35)

Contrary to what is hypothesized based on the theoretical framework that is presented in chapter 2, the results of the analyses in model 1, 2 and 3 indicate that the use of different dimensions of performance does not result in differences in the direction of organizational change in response to performance feedback. In order to find out if firm size has any moderating effects on the differences in the direction of change between the different dimensions of performance on the relationship between performance feedback and acquisition behavior, the interaction of firm size and performance feedback is included in model 4. Because of the insignificant results that are found for performance feedback using both growth and market share as a dimension of performance, model 4 is performed only for acquisition count and performance feedback focusing on ROA and Tobin’s Q as performance dimensions.

The interaction terms are generated by separately multiplying performance below and above aspirations by firm size, for both performance feedback using growth as a dimension of performance and performance feedback using Tobin’s Q as a dimension of performance. The results of the regression analysis incorporating the interaction of firm size and performance feedback are presented as model 4 in table 3 and table 4. Table 3 shows that, for performance feedback using ROA as a dimension of performance, the results for model 4 (.832, p < .001) are positive and significant in case of performance above the aspiration level. In other words, acquisition count increases as performance rises above the aspiration level. This is in contrast with hypothesis 1. As performance falls below the aspiration level, table 3 shows that the results for model 4 (-.407, p < .001) are negative and significant. Figure 3 provides an illustration of the results of the regression analysis for performance feedback using ROA as a dimension of performance and acquisition count, incorporating the moderating effect of the interaction of firm size and performance feedback. The figure shows that, as performance falls below the aspiration level, acquisition count increases. These results are against hypothesis 2, which states

(36)

that, when using performance measured by ROA, acquisition count decreases as performance falls below the aspiration level.

Figure 3

Regression results for performance feedback (ROA) and acquisition behavior: Controlling for the interaction of firm size and performance feedback

The results that are presented in table 4 show that, for performance feedback using Tobin’s Q as a dimension of performance, the results for model 4 (-.0426, p < .001) are negative and significant in case of performance above the aspiration level. This indicates that acquisition count decreases as performance rises above the aspiration level, supporting hypothesis 1. As performance falls below the aspiration level, table 4 shows that the results for model 4 (.0469, p < .001) are positive and significant. The results of the regression analysis for performance feedback using Tobin’s Q as a dimension of performance and acquisition count, incorporating the moderating effect of the interaction of firm size and performance feedback, are illustrated in figure 4. As can be seen in the figure, acquisition count decreases as performance falls below the aspiration level. This is against hypothesis 3, stating that acquisition count increases as a response to performance below the aspiration level, when focusing on Tobin’s Q as a dimension

obtained results (ROA)

(37)

Figure 4

Regression results for performance feedback (Tobin’s Q) and acquisition behavior: Controlling for the interaction of firm size and performance feedback

In sum, the regression analyses have led to the following outcomes with regards to the hypotheses that are developed in chapter 2: First, performance feedback using ROA and Tobin’s Q are significantly related to acquisition count. The results for acquisition count and performance feedback using growth and market share, however, were not significant. Second, all of the models in the analysis, except for model 4 focusing on Tobin’s Q as a dimension of performance, have rejected hypothesis 1. Third, hypothesis 2 is supported by model 2 and 3, and is rejected by model 4 focusing on ROA as a dimension of performance. At last, hypothesis 3 is rejected by all of the models in the analyses.

obtained results (Tobin's Q) hypothesized results acquis it ion c oun t

(38)

V. DISCUSSION AND CONCLUSION

The primary objective of this study was to provide and test an explanation for the conflicting results in the literature on performance feedback and organizational behavior. While the Behavioral Theory of the Firm by Cyert & March (1963) predicts that performance below the aspiration level causes firms to induce a process of problemistic search, resulting in an increase in organizational change, studies investigating this relationship have found contradictory results (Audia & Greve, 2006; Iyer & Miller, 2008). Building on the Attention-Based View of the Firm by Ocasio (1997), research in the field of performance feedback theory seems to have overlooked the fact that organizations may interpret strategic actions differently (Greve, 2011). Decision makers are constrained in the process of decision-making by bounded rationality and limited managerial attention (Ocasio, 1997; Greve, 2003a). In order to deal with these constraints, they base their decisions on learning from performance feedback, i.e. searching and evaluating alternatives until the one that fulfills the performance goal or aspiration level has been found (Greve, 2003a). Therefore, managerial attention is channelized towards strategic actions that contribute to obtaining their performance goals. Since aspiration levels often reflect one single dimension of organizational performance, the performance dimension that is used in establishing the aspiration levels specifies the performance goal towards which managerial attention is directed (Bromiley & Harris, 2014).

Combining the theoretical perspectives that are derived from the Behavioral Theory of the Firm (Cyert & March, 1963) and the Attention-Based View of the Firm (Ocasio, 1997), this study proposes the notion of differences in the dimensionality of measures for performance as a new theoretical explanation for the conflicting results that are found in performance feedback literature. Building on the results that are found by Combs et al. (2005), ROA, Tobin’s Q, growth, and market share are introduced as examples of the different dimensions of performance. In order to determine the effects of the different dimensions of performance on