• No results found

The impact of ambidexterity on firms’ short- and long-term performance

N/A
N/A
Protected

Academic year: 2021

Share "The impact of ambidexterity on firms’ short- and long-term performance"

Copied!
36
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The impact of ambidexterity on firms’ short- and long-term performance

EPMS University of Amsterdam

Executive Programme in Management Studies – Strategy Track

Student : Julia IJsselhof

Student number : 10901698

Primary supervisor : Bernardo Silveira Barbosa Correia Lima

Version : Final

(2)

1

Statement of Originality

This document is written by Student Julia IJsselhof 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)

2 Table of contents

Abstract...3

1. Introduction ...4

2. Literature review...7

2.1 The (potential) effect of ambidexterity on performance...9

2.2 Environmental moderators on the ambidexterity–performance relationship...11

2.2.1. Environmental dynamism as a moderator...11

2.2.2. Environmental munificence as a moderator...13

3. Data and method...14

3.1 Dataset………14

3.2 Variables……….16

3.3 Analysis methods...19

3.4 Assumption check………...21

4. Results...22

4.1 Impact of ambidexterity on firms’ short-term performance...23

4.2 Impact of ambidexterity on firms’ long-term performance...26

5. Discussion...28

5.1 The effect of ambidexterity on firms’ short-term performance...28

5.2 The effect of ambidexterity on firms’ long-term performance...30

5.3 Limitations and further research……….30

6. Conclusion...31

(4)

3 Abstract

This article seeks to clarify the effect of ambidexterity on firms’ short- and long-term performance. Ambidexterity refers to the capability of a firm to simultaneously exploit and explore. According to the ambidexterity premise, firms that are capable of building on efficiency, while simultaneously renewing, will encounter a higher performance. However, the empirical results are ambiguous.

In this research, we tested the effect of ambidexterity on three performance dimensions: Return on assets (ROA), Tobin’s Q and Growth. The former is a short-term performance indicator whereas the latter two are long-term performance indicators. Here, we tested whether exploitation and exploration are not acting independently, but positively reinforce each other. In addition, we tested whether environmental dynamism and munificence moderate this relationship.

Our findings suggest that ambidexterity positively contributes to firms’ short-term performance (i.e. ROA). In the short term, this effect is more pronounced for firms in low dynamic environments compared to firms in high dynamic environments. We found that ambidexterity also positively influences one of the long-term performance indicators (i.e. Tobin’s Q), but ambidexterity does not significantly impact the other long-term performance indicator (i.e. Growth). In addition, in both the short and long term, ambidextrous firms in high munificent environments outperform ambidextrous firms in low munificent environments.

(5)

4 1. Introduction

There is a lot of attention in literature towards the causal relationship between ambidexterity and performance. Ambidexterity means that firms simultaneously engage in exploitation as well as exploration. Exploitation is related to efficiency and existing competencies and markets. On the other hand, exploration is related to innovation and the renewing of competencies and markets (Gupta, Smith & Shally, 2006). The ambidexterity premise suggests that firms that are capable of simultaneously pursuing exploration and exploitation will have an overall better performance compared to firms that emphasize one activity at the expense of the other (Tushman & O’Reilly,1996). Ambidexterity would enhance performance because it enables the firm to be flexible and innovative without derogating the routines and stability (Simsek, 2009). The optimal mix of exploitation and exploration will yield firm performance because it ensures both the short- and long-term success of the firm (March, 1991; Junni, Sarala, Taras & Tarba, 2013). Studies in favour of ambidexterity state that ambidexterity has a positive effect on the financial performance of the business unit (Jansen, Van den Bosch & Volberda, 2006), on sales growth rate of the firm (He & Wong, 2004) and towards the market value of the firm (Uotila, Maula & Zahra, 2009). However, opponents of ambidexterity state that ambidexterity negatively influences firm performance. These studies found the opposite effect: For example, Wernerfelt and Montgomery (1988) state that firms which pursue both exploitation as well as exploration are at risk of being mediocre at both, because they sacrifice internal consistency compared to more focused firms. According to Venkatraman, Lee & Iyer (2007), sequential ambidexterity, in contrast to simultaneous forms of ambidexterity, is a more significant predictor of sales growth. In addition, according to Atuahene–Gima (2005), the interaction between competence exploitation and exploration negatively impacts radical innovation performance. As may be clear, the findings regarding the effect of ambidexterity on performance are ambiguous.

(6)

5 Besides these contradicting findings, different moderators may influence the strength and form of the relation between two variables and therefore they might explain the different findings (MacKinnon, D. P., 2011). These moderators are important to consider, because the success of ambidexterity may be contingent of environmental aspects. This means that ambidexterity may lead to different outcomes under different contextual conditions (Jansen et al., 2006). These moderators can be diverse; they may vary from organizational aspects to environmental conditions. Several studies have focused on these environmental factors that may moderate the relation between ambidexterity and performance. For example, Uotila et al. (2009) showed that the optimal balance between exploration and exploitation depends on environmental conditions. According to Jansen et al. (2006), explorative innovation is more beneficial in dynamic environments. In addition, Wang and Li (2008) state that the harm of overexploitation is more pronounced in high dynamic environments. They also showed that the negative effect of overexploration is weaker in high dynamic environments.

There may be different possible explanations for these conflicting findings regarding the question whether ambidexterity positively influences firm performance. It may be possible that several moderators play a role in helping to explain the conflicting findings of the ambidexterity–performance relationship (Raisch, & Birkinshaw, 2008). This means that the effect of ambidexterity on performance is either stronger or weaker when the moderator is present.

It might also be possible that ambidexterity positively influences some types of performance while at the same time negatively influencing other types of performance. It is important to consider the different types of performance, because the question whether a firm is successful depends on how performance is measured. A firm can be labeled as successful according to

(7)

6 one of the interpretation of performance, while it might not be according to the standards of another (Kirby, 2005). In addition, the conflicting findings may also come from methodological choices, different conceptualizations of the concept and their research design (Junni et al. 2013).

The aim of this thesis is to contribute to a more fine-grained view on the ambidexterity– performance relationship. This will be done by analyzing the effects of ambidexterity on three types of performance: ‘Accounting return’, ‘stock market measures’ and ‘growth measures’ (Combs, Crook & Shook, 2005). Accounting returns are short-term performance indicators, whereas stock market measures and growth measures are long-term performance indicators. By doing so, we would like to investigate whether ambidexterity may have a different impact on the different types of performance. We would also like to discover whether environmental factors may influence this ambidexterity–performance relationship. We will test the moderators as amplifiers. By examining the effect of ambidexterity on the different types of performance, while at the same time including environmental moderators, we hope this research will lead to a more complete and holistic view of ambidexterity.

This paper is structured as follows: In ‘Chapter 2’, we will provide a review of the existing literature and describe the lack of agreement among researchers regarding the performance implications of ambidexterity. Here we describe our hypotheses. In ‘Chapter 3’ there is attention towards the data and methods by which we will answer our hypotheses. In ‘Chapter 4’, we will present the results. These results are further discussed in ‘Chapter 5’. Here we will describe how this research helps to resolve the conflicting findings currently present in the literature. Finally, we will end with the conclusions in ‘Chapter 6’.

(8)

7 2. Literature review

Nowadays, companies face intensive competition and today’s activities may be obsolete tomorrow (e.g. Derfus, Maggitti, Grimm & Smith, 2008; Prahalad & Hamel, 1990). Therefore, companies need to renew themselves (Agarwal & Helfat, 2009). A company can have different drivers to invest in research and development. This can come from either bad performance that causes ‘problemistic search’, or come from the excess of slack resources that are earned from good performance that cause ‘slack search’ (Greve, 2003; Chen, 2008). In addition, a company may renew due to declining markets or major changes in customer demands (Agarwal & Helfat, 2009).

Because companies should secure both short-term goals and long-term organizational survival, companies are forced to pay attention to both today’s situation as well as to the future and renewal of the company. Firms can fulfil these, sometimes conflicting, goals by pursuing ambidexterity. As mentioned earlier, ambidexterity means that firms simultaneously engage in both exploitation as well as exploration. According to Levinthal and March (1993), exploitation ensures the current viability of a company, while exploration ensures its future viability. Ambidexterity does not lead to radical changes, but instead firms make small changes to remain a fit with the environment. An ambidextrous firm needs to balance both exploitation and exploration and those two activities need to be above a certain level. A situation where a company has low levels of exploitation and exploration can be seen as balanced, but in this case there is an absence of ambidexterity (He & Wong, 2004; Simsek, 2009). The appropriate balance between exploitation and exploration does not need to be 50/50, but may depend on environmental conditions (Uotila et al 2009) and firm specific characteristics (Wang &.Li, 2008). Ambidextrous firms are thus firms that simultaneously execute exploitative activities as well as explorative activities. Non-ambidextrous firms are firms that focus solely or mainly on either exploitation or exploration, or have a low level of

(9)

8 both.

There are conflicting findings regarding the question whether ambidexterity positively influences firm performance. Besides these conflicting findings, the existing studies of the ambidexterity–performance relationship, also vary widely at some crucial point. As a consequence, the different perspectives on organizational ambidexterity lead to disconnectedness and complexity of the field (Raisch, & Birkinshaw, 2008). We will discuss six of the key issues that contribute to the complexity of the field below.

Firstly, there is a lack of consensus in the literature about the relation between exploitation and exploration. Some consider them to be two ends of a continuum, where others think they are orthogonal to each other, meaning that they are two different things (Gupta et al., 2006). In this study, we consider ambidexterity as being orthogonal. Secondly, the units of analysis differ; some researchers focused on the firm level (Katila & Ahuja, 2002), whereas others focused on the business unit (Jansen et al. 2006) or the individual level. The level of analyses in this study relates to the firm as a whole. Thirdly, there are different views on the mode for managing conflicting pressures of exploitation and exploration. One of those is by integration. This is called ‘contextual ambidexterity’. The remaining three modes for managing the conflicting pressures are by separation: Organizational separation, temporal separation and domain separation (Lavie, Stettner & Tushman, 2010). This issue is not related enough to the research question to pay further attention to in this research. Fourthly, there is lack of agreement on how to measure and define performance. The question how to define which firm outperforms others, depends on the timeframe and the definition of performance. Kirby (2005) questions this as follows: “Are the winners the ones with the highest market caps, the ones with the greatest sales growth, or simply the ones that remain standing at the end of the game? (And when is the end of the game?)”. When performance is measured in terms of firm survival, then the firm’s ability to adapt to its environment is key. However, when a company

(10)

9 adapts to the current environment, this will reduce the company’s capacity to adapt to future changes. This is called the ‘adaptation paradox’ (Burgelman, 1991). In this research we differentiate between accounting returns measures, stock market measures and growth measures. We chose to test the effect of ambidexterity on these performance measures, because according to Combs et al. (2005) organizational performance consists of these three dimensions. We therefore decided to measure the effect of ambidexterity on all three performance dimensions. Fifthly, some of the studies acknowledged the role of moderators that may have an impact on the relation between ambidexterity and performance. The nature of these moderators vary from internal mechanisms (Jansen et al., 2006) to external factors like the R&D intensity of the industry (Uotila et al., 2009). Sixthly, although both ambidexterity and punctuated equilibrium have the goal to remain a fit with the environment, they have different routes to accomplish this fit. By pursuing ambidexterity, firms continuously make small changes to remain this fit. Punctuated equilibrium, in contrast, is an approach by which, after long periods of stability, firms radically change in a short time (Romanelli & Tushman, 1994). This paper focuses on ambidexterity.

Although all these key issues are important, the main focus of this research is on how to measure and define ambidexterity (the firstly mentioned key issue), the question how to measure and define performance (key issue four) and whether and how environmental factors may moderate the relationship between ambidexterity and performance (key issue five).

2.1 The (potential) effect of ambidexterity on performance

As said, in this research, ambidexterity is perceived as orthogonal, meaning that exploitation and exploration are two different things. If ambidexterity is perceived as orthogonal, the correct test for performance implications would be to test for a positive interaction effect between exploitation and exploration on performance (Gupta et al., 2006). This means that we

(11)

10 will test whether the simultaneous pursuit of both activities adds or detracts from each other’s value. We thus test whether the two factors interact, meaning whether the relationship between one predictor and the dependent variable changes depending on another predictor (Cohen, Cohen, West & Aiken, 2003). By doing this research, we investigate whether ambidextrous firms have a better performance compared to non-ambidextrous firms. Researchers in favour of ambidexterity state that exploitation and exploration are not independent, but positively reinforce each other (He & Wong, 2004). For example, according to Katila and Ahuja (2002), a precondition of exploration is that the firm exploits the existing capabilities. Conversely, the exploration of new capabilities is needed in order to enhance the existing knowledge base of the firm. According to Tushman, Smith, Wood, Westerman & O’Reilly (2003), firms which do not have ambidextrous designs are unable to implement innovation streams. In addition, ambidextrous firms may perform better compared to non-ambidextrous firms, because they have the capability to simultaneously exploit and explore. Therefore, they are less prone to falling in the ‘myopia trap’ and the ‘failure trap’. Due to myopia, a company can pay too much attention towards exploitation. This is what is known as the myopia trap (Levinthal & March, 1993). On the other hand, a company can pay too much attention towards exploration, for instance when a company tries to solve a failed initiative with more explorative efforts, which might also fail. This is referred to as the failure trap (Levinthal & March, 1993). When we take the arguments above into account, this leads us to the following hypotheses:

Hypothesis 1a: Ambidextrous firms perform better on the short term compared to non-ambidextrous firms.

Hypothesis 1b: Ambidextrous firms perform better on the long term compared to non-ambidextrous firms.

(12)

11

Figure 1. Conceptual model of hypotheses 1a and 1b

2.2 Environmental moderators on the ambidexterity–performance relationship

We also expect that environmental factors may play a role in explaining the contradictory findings in the existing literature. Dess and Beard (1984) identified three environmental dimensions: Dynamism, munificence and complexity. Environmental dynamism relates to the question how fast and unpredictably the environment is changing (Simsek, 2009; Jansen et al. 2006; Wang & Li. 2008). Environmental munificence relates to how much growth potential an environment has. Environmental complexity relates to how complex and heterogeneous the company’s task environment is (Dess & Beard, 1984). In this study, we will focus on environmental dynamism and munificence as (potential) moderators.

2.2.1. Environmental dynamism as a moderator

Environmental dynamism is directly related to the question how fast the company’s products and services are obsolete. Companies that act in high dynamic environments have products that are prone to become obsolete in the short term. The more dynamic the environment, the less a firm can rely on past knowledge and the more the company needs to explore (Zahra & George, 2002). On the contrary, firms that act in low dynamic environments are less prone to their products and/or services becoming obsolete. In low environmental dynamism conditions, firms have less pressure to explore and therefore can focus more on exploitation. In these

(13)

12 environments, the environmental conditions change slowly, which gives the firm the possibility to 1) continue their current business, or 2) imitate the inventions of its competitors, which is less costly than inventing it itself. Because these firms can invest a greater percentages of their financial resources in exploitation, we expect that these firms have a higher short-term performance (e.g. the return on assets, or ROA) compared to firms in high dynamic environments.

The more dynamic the environment, the quicker the product is obsolete and therefore the greater the pressure for firms to invest in exploration (Uotila et al. 2009). In these dynamic environments, firms are forced to simultaneously pay attention to both today’s situation as well as the future (Birkinshaw & Gibson 2004). Firms in high dynamic situations cannot remain in the same place or simply imitate their competitors. They need to quickly come with a new idea for itself and these firms need to rely on their own strengths. This is one of the key principles of the dynamic capability view (Teece, Pisano & Shuen, 1997). As a consequence, these firms need to invest a greater percentage of their resources in exploration, which will negatively influence the short term. This leads us to hypothesis 2a:

Hypothesis 2a: Ambidextrous firms in low dynamic environments have a higher short-term performance compared to ambidextrous firms in high dynamic environments.

‘Figure 2’ represents the conceptual model of hypothesis 2a.

(14)

13 2.2.2. Environmental munificence as a moderator

High munificent environments are characterized by its high growth potential. In these high munificent conditions, there is enough place for the firm and its competitors. Here, the rival’s actions do not negatively influence the initial firm performance (Derfus et al., 2008). In contrast, low munificent environments have less growth potential. In these environments, competition is intensified. In low munificent environments, firms are hindered by its competitors and the rival actions may negatively influence the focal firm performance (Derfus et al, 2008). According to Cao, Gedajlovic & Zhang (2009), in high munificent environments, firms have access to more external resources, which can help the firm to attenuate resource constraints. In these conditions, firms more easily get the resources that are necessary to pursue ambidexterity. On the contrary, in less munificent environments, firms will have greater difficulty in obtaining the necessity resources. We therefore expect that firms in low munificent environments perform less optimal, both on the short-term as on the long-term performance, compared to firms in high munificence environments. This leads us to the last hypotheses:

Hypothesis 2b: Ambidextrous firms in high munificent environments perform better on the short term compared to ambidextrous firms in low munificent environments.

Hypothesis 2c: Ambidextrous firms in high munificent environments perform better on the long term compared to ambidextrous firms in low munificent environments.

(15)

14 In ‘Figure 3’, the conceptual model of hypotheses 2b and 2c is presented.

Figure 3. Conceptual model of hypotheses 2b and 2c.

3. Data and method

3.1 Dataset

This research was done by analyzing secondary data. In particular, by analyzing the data of Compustat, which is a database which contains information of all firms traded in the United States’ stock market. The dataset contains data of firms in the following industries: ‘Manufacturing: Chemicals and allied products (SIC 28)’, ‘Manufacturing: Industrial and commercial machinery and computer equipment (SIC 35)’ and finally ‘Manufacturing: Electronics and other electrical equipment (SIC 36)’. These companies were merged with patent citation data from the NBER patent citations file. The original dataset contained 58,522 observations, coming from 4,572 firms within the time period of 1970 through 2008.

Our approach is novel in the way that we used the same dataset to test the effect of ambidexterity on different types of performance, which makes the results comparable. Therefore it was preferable for all the firms in the sample to contribute to all the regression tests. Otherwise we would have included, for example, company A, for investigating the effect of ambidexterity on ROA, while this same company would not have been included for measuring the effect of ambidexterity on Growth. Consequently, we decided to use a

(16)

15 restriction that excludes firms that contain one or more missing variables, and that also excludes firms that contain less than 500 employees, when executing the regression analysis. We did this because large firms may have more patents than small firms (Wang & Li, 2008). We did not standardize or mean centre the variables, because in this research the mean is not a logical zero point; we want to avoid the situation where in the interaction term, a mean score on exploitation or exploitation causes the outcome of the interaction to be zero. In our case, it would be preferable if a very low score on either exploitation or exploration, or both, causes the outcome of the interaction to be zero. Therefore, we chose not to standardize or mean centre the variables.

In this research, we tested whether ambidextrous firms differ from non-ambidextrous firms in their performance. We thus wanted to decrease the risk that there may have been other explanations for the difference in performance between ambidextrous and non-ambidextrous firms. Some of these alternative explanations were ruled out by our sample choice, in the following ways. First of all, we chose to analyze solely manufacturing firms which are traded in the U.S. stock market, so that differences in industries and differences in national technological characteristics could not interfere with outcomes (Katila & Ahuja, 2002). In addition we excluded the alternative explanation that the different findings come from the propensity for patenting that may vary considerably across industries, as all the firms were manufacturing firms.

Exploitation and exploration efforts do not pay off immediately. We therefore decided to lag the dependent variables by the period of one year. Because of the multilevel nature of our data (multiple firms from multiple years), we needed to decide whether to test for a fixed or for a random effect. We used the Hausman test to help make this decision. The null hypothesis of

(17)

16 this test is that random effects are preferred, the alternative is that a fixed effects are favorable. In this research, we conduct the regression test nine times in total, due to the different moderators and dependent variables. In eight out of nine tests, the Hausman test was significant, which indicates that the fixed effect was more appropriate. The remaining test was not significant, which indicates that the random effect would be favorable; this was the simple interaction test between ambidexterity and growth. We performed the test accordingly.

3.2 Variables

a) Dependent variables

In this research, we followed the argumentation of Combs et al. (2005) that organizational performance consists of three related but distinct dimensions: Accounting returns, stock market and growth measures. These three dimensions are distinct from each other in the way that there is agreement that the constructs within these dimensions relate to each other (this is known as convergent validity) and that there is discriminant validity, meaning that the measurements of different constructs do not converge (Combs et al. 2005). We decided to measure the effect of ambidexterity on these three performance measurements; the effects of ambidexterity on accounting returns (measured as ROA), stock market measurement (measured as Tobin’s Q) and growth measurement (measured as Growth).

ROA

We chose ROA as a proxy for the accounting return measurement. As mentioned earlier, the abbreviation ROA stands for ‘Return On Assets’. We define ROA as a short-term performance indicator. ROA is an appropriate indicator to measure short-term performance, because ROA is focused on past performance and thus is retrospective in nature (Row & Morrow, 1999). ROA is a profitability ratio that reveals how efficient a company is using its assets in generating earnings. We can calculate the ROA by dividing a company’s annual net

(18)

17 income by its total assets. By using the ROA we can compare different companies in different environmental conditions and get insight in how efficient they generate earnings. Different industries may demand different levels of capital requirements (e.g. Porter, 1979). However, this sample only contains manufacturing firms. We decided to test for ROA before income tax, because different industries may have to pay different income tax percentages.

Tobin’s Q

We chose Tobin’s Q as a proxy for the stock market measurement and define Tobin’s Q as a long-term performance indicator. The reason why Tobin’s Q is a long-term performance indicator, is because it is forward looking (Wang & Li, 2008). Stock market measurements reflect the expectations of society on the firms potential to create shareholder value (Row & Morrow 1999). By using Tobin’s Q, we are thus able to examine performance effects over varying time horizons. We can calculate the Tobin’s Q as the market value of assets divided by the book value of assets (Uotila et al., 2009).

Growth

The growth measurement dimension is represented by growth. Growth is a long-term performance indicator, because it can only be measured by comparing firm performance in a certain period to firm performance in another period. According to Junni et al. (2013), we can distinguish between absolute and relative performance. If we apply this on the growth dimension we could look at the absolute performance in raw growth numbers. We could also compare the company’s performance relative to that of its competitors, and look at relative market share. In this research we operationalized growth by the company’s absolute performance.

(19)

18 b) Independent variables

In this research, we built on Katila and Ahuja (2002) in operationalizing exploitation and exploration. They operationalized these two concepts by means of patent data. They called exploitation ‘search depth’, and measured this construct as the average number a firm repeatedly used the citations in the patents it applied for. When an existing citation is used, this means that the firm uses its current knowledge base, and search depth therefore represents exploitation. Katila and Ahuja (2002) called exploration ‘search scope’ and measure this as new patent citations as a percentages of all the patent citations of a firm in a certain year. A citation is perceived as new if the citation cannot be found in the previous five year list of patents and citations (Katila & Ahuja, 2002). When a company applies for a new citation, this means the firm is creating a new knowledge base, and search scope therefore represents exploration.

We measured search scope with the exact same criteria as Katila and Ahuja (2002). However, instead of measuring search depth by comparing new citations only to the focal year total citations, we measured depth by contrasting a firm’s average number it uses the patents, to the patents a firm applied for the last five years. Both depth and scope are at a scale of ‘0’ to ‘1’.

c) Moderating variables

In this study, we tested environmental dynamism and munificence as potential moderators. We follow on Boyd (1995) by measuring environmental dynamism as the environmental volatility. This environmental dynamism is measured by using a measure of the volatility of industry sales growth rate over the same period of time (Boyd, 1995). Munificence was operationalized as the abundance of resources. Munificence was also operationalized in line with Boyd (1995) by using a measure of industry sales growth over a five year period. Both

(20)

19 environmental dynamism and munificence are measured as a continuum.

d) Control variables

Because scope and depth may be influenced by firm size, we decided to control for firm size. We operationalized firm size as the number of employees and added this as a control variable in our regression formula. In eight out of nine tests, we used the fixed effect approach. This means that in these tests we therefore controlled for time invariant variables.

3.3 Analysis methods

As mentioned earlier, our analysis consists of nine regressions, of which eight are fixed effects tests and one is a random effect test. The independent variables depth and scope are used to predict the performance of the firms. Because we have three performance indicators, three regression tests were performed. Furthermore, we tested for two moderating variables, resulting in another six analyses. Both types of questions about the ambidexterity and moderator effects can be answered by analyzing the interaction components of the regression. In the following section, this is described in more detail.

The regression formula for a regression analysis with two predictors and a simple interaction is: y = β0 + β1X1 + β2X2 + β3X1X2 + e

Variable y is the outcome variable ROA, Tobin’s Q or Growth. X1 and X2 are the two

independent variables depth and scope. In case of ambidexterity, both X1 (depth) and X2

(scope) are non-zero. It is only in this situation that the interaction term in the regression formula becomes important. In all other situations, one part of the interaction term is zero (or very close to zero) resulting in the interaction term being zero as well. Consequently, by analysing the beta coefficient of the interaction term, we can answer hypotheses 1a and 1b:

(21)

20 ‘Ambidextrous firms perform better on the short and long term compared to non ambidextrous firms’.

The model for these hypotheses is shown in ‘Figure 4’, below. The arrow going from X1 to y

refers to beta coefficient 1, the arrow from X2 to y refers to beta coefficient 2, and, very

important, the arrow going from X2 to the effect of X1 on y, refers to the interaction effect

(beta coefficient 3). By conceptualising the regression like this, it is easy to see that ambidexterity is at hand only in the latter arrow; both X1 and X2 must have an effect on y for

ambidexterity to be at hand.

Now, if we would like to include the environmental moderators and test hypotheses 2a up to 2c, we need to test for a three way interaction, because the test for a positive interaction effect between depth and scope is already an interaction itself. This would be an interaction of an interaction. The model of this is pictured in ‘Figure 5’. The accessory formula is as follows: y = β0 + β1X + β2Z + β3W + β4XZ + β5XW + β6ZW + β7XZW + e

In ‘Figure 5’, the blue arrow refers to the three way interaction: It is an effect of X3 on the

effect of X2 on the effect of X1 on y. Accessory beta coefficient is number 7.

(22)

21 3.4 Assumption check

Before performing the analysis, we checked for linear relations between the dependent and the independent variables. All these relations had an approximately linear distribution. When we investigated the existence of possible outliers, we discovered that there were a lot of extreme values. These ‘outliers’ were present in such large numbers that one can argue that they may not be seen as unusual. Therefore, we did not delete these ‘extreme’ observations. We also checked whether the errors were approximately normally distributed by executing the Jarque-Bera (JB) test for normality. The 0 hypothesis of the Jarque–Bera test claims that the errors are normally distributed. In our tests, JB>Chi2 and therefore the 0 hypothesis was rejected. This means that the errors are not normally distributed. In order to deal with this problem, we included the robustness option (also called ‘Huber/White/Sandwich’ estimator) in our regression analyses. By doing so, the concerns regarding the errors are reduced.

(23)

22 4. Results

In this chapter, we discuss the data analyses and our findings. ‘Table 1’ below reports these findings. Results of regression analyses for the different performance indicators

Variable Model 1 ROA Model 2 Tobin’s Q Model 3 Growth Model 4 ROA Model 5 Tobin’s Q Model 6 Growth Model 7 ROA Model 8 Tobin’s Q Model 9 Growth Intercept .08 *** 1.81 *** .29 *** .10 *** 1.96 *** .25 *** .06 *** 1.72 *** .08 * Number of employees .00 .00 .00 .00 .00 .00 .00 .00 .00 Depth -.08 *** -.09 -.20 .03 .37 .12 -.10 *** -.76 -.27 Scope .01 * .00 .05 .01 ** .03 .06 -.01 * -.10 -.03 Depth x Scope .17 *** 1.51 * .79 -.04 .28 -.79 .21 *** 3.20 *** 1.35 * Dynamism -.77 *** -7.43 *** -4.18 ** Dynamism x Scope -.32 -1.30 -.42 Dynamism x Depth -5.31 ** -19.96 -18.77

Dynamism x Scope x Depth 9.31 ** 54.86 82.37

Munificence .37 *** 1.58 *** 1.66 ***

Munificence x Scope .20 *** 1.06 0.76

Munificence x Depth 1.33 *** 17.94 ** 1.39

Munificence x Scope x Depth -2.20 *** -38.24 *** -10.13 **

(24)

23 ‘Table 1’ shows the results of the regression analyses. Four of our five hypotheses are supported. The prediction in hypothesis 1b is not fully supported, since we found that ambidexterity positively influences one of the long-term performance indicators (Tobin’s Q) but ambidexterity does not significantly impact the other long-term performance indicator (Growth). We found that ambidexterity (i.e. the interaction between depth and scope) significantly predicts firms’ ROA. This finding supports our hypothesis 1a, which states that ambidexterity positively impacts short-term performance. Moreover, this short-term performance is significantly moderated by environmental dynamism, which is predicted in hypothesis 2a. In addition, this short-term performance is also significantly moderated by munificence, which is predicted in hypothesis 2b. Furthermore, the tests indicated that environmental dynamism does not moderate the effect of ambidexterity on the long-term performance of the firm. However, when we test for munificence as a moderator for the effect of ambidexterity on firms’ long-term performance, it appears that munificence significantly does. Therefore hypothesis 2c is supported.

‘Table 1’ does not inform about the direction and slope of the effects. To gain further insights into these effects, we added plots of the significant interactions to this chapter. As described earlier, depth represents exploitation whereas scope represents exploration.

4.1 Impact of ambidexterity on firms’ short-term performance

We found that there is a significant interaction effect between depth and scope on ROA. It thus appears that the effect of depth on ROA depends on the level of scope and vice versa. As can be seen in ‘Figure 6’, depth has a positive effect on ROA when scope is high, but has a negative effect on ROA when scope is low. This figure shows that firms which simultaneously pursue exploitation and exploration have a higher ROA compared to

(25)

non-24 ambidextrous firms. Summarizing, we can say that ambidextrous firms are better off in the short term (when we look at the ROA indicator) compared to non-ambidextrous firms. This result therefore confirms hypothesis 1a, and therefore this hypothesis is accepted.

This research is also concerned with the question whether the effect of ambidexterity on firms’ short-term performance is influenced by environmental dynamism and munificence. We found that both environmental dynamism and munificence are significantly influencing the effect of ambidexterity on firms’ short-term performance. As can be seen in ‘Figure 7’, ambidextrous firms in low dynamic environments have a higher short-term performance compared to ambidextrous firms in high dynamic environments. Consequently, we accept hypothesis 2a. In addition, this figure reveals that ambidextrous firms in both low and high dynamic environments, outperform the non-ambidextrous firms which are positioned in the same environment.

Also, munificence appears to influence the effect of ambidexterity on the short-term performance. As can be seen in ‘Figure 8’, ambidextrous firms that act in high munificent environments have a better short-term performance compared to ambidextrous firms in low munificent environments. The ROA of firms in low dynamic environments is negative, whereas the ROA of firms in high dynamic environments is positive. We therefore accept hypothesis 2b.

(26)

25 -0,010 -0,005 0,000 0,005 0,010 0,015 0,020

depth low depth high

R OA ROA scope low scope high -0,05 -0,04 -0,03 -0,02 -0,01 0,00 0,01 0,02

depth low depth high

ROA

Dynamism

scope high & dynamism high scope high & dynamism low scope low & dynamism high scope low & dynamism low -0,02 -0,01 0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07

depth low depth high

Roa

Munificence

scope high & munificence high scope high & munificence low scope low & munificence high scope low & munificence low

Figure 6. Interaction effect on ROA

Figure 7. Influence of dynamism on ROA

(27)

26 4.2 Impact of ambidexterity on firms’ long-term performance

The results visualized in ‘Figure 9’ indicate that ambidexterity significantly impacts the Tobin’s Q ratio of the firm. When both depth and scope are high, this leads to a high Tobin’s Q ratio as well. When depth and/or scope is low, this leads to a low, and sometimes even negative, Tobin’s Q ratio. However, ambidexterity appears to not significantly influence the growth of the firm. We can thus say that ambidexterity influences one of the long-term performance indicators, while at the same time it does not significantly impact the other long-term performance indicator. Therefore, we cannot accept or reject hypothesis 1b.

We decided to also test whether environmental dynamism and munificence may influence long-term performance. We found that munificence significantly moderates the impact of ambidexterity on firms’ long-term performance. Ambidextrous firms that act in high munificent environments generally have a better Tobin’s Q ratio compared to ambidextrous firms that act in low munificence environments. These results are visualized in ‘Figure 10’. In addition, firms in low munificence environments have a better Tobin’s Q ratio if they invest in both depth and scope, instead of being non-ambidextrous. However, surprisingly, firms in high munificent environments that invest in depth but do not invest in scope have a better Tobin’s Q ratio compared to ambidextrous firms. We can thus conclude that in high munificent environments, ambidexterity leads to a suboptimal long-term performance when we look at the Tobin’s Q ratio.

Munificence also significantly influences the effect of ambidexterity on Growth. This is visualized in ‘Figure 11’. Ambidextrous firms in high munificent environments have a higher growth ratio compared to ambidextrous firms in low munificent environments. We could therefore accept hypothesis 2c. Ambidextrous firms in high munificent environments perform better, in the long term, compared to ambidextrous firms in low munificent conditions.

(28)

27 -0,05 0,00 0,05 0,10 0,15

depth low depth high

To b in 's Q Tobin's Q scope low scope high -0,2 -0,1 0,0 0,1 0,2 0,3 0,4

depth low depth high

To b in 's Q Munificence

scope high & munificence high

scope high & munificence low

scope low & munificence high

scope low & munificence low -0,10 -0,05 0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35

depth low depth high

Gr

o

wt

h

Munificence

scope high & munificence high scope high & munificence low scope low & munificence high scope low & munificence low

Figure 9. Interaction effect on Tobin’s Q

Figure 10. Influence of munificence on Tobin’s Q

(29)

28 5. Discussion

We will first focus on the explanation of the impact of ambidexterity on the short-term performance of the firm in paragraph 5.1. Subsequently, we will describe the impact of ambidexterity on the long-term performance of the firm in paragraph 5.2.

5.1 The effect of ambidexterity on firms’ short-term performance

The results suggest that ambidextrous firms are better off compared to non-ambidextrous firms in the short term. This could be caused by depth and scope not acting independently; they instead positively reinforce each other. This is in line with previous research conducted by He and Wong (2004) and Katila and Ahuja, (2002), because they too found a positive interaction effect. These findings may also be explained by the underlying reasons why firms explore. As is discussed earlier, a company can have different drivers to invest in exploration. This can come from either bad performance which causes ‘problemistic search’, or come from the excess of slack resources that are earned from good performance which causes ‘slack search’ (Greve, 2003; Chen, 2008). The latter explanation for exploration may explain our findings. Firms which perform well on the short term may have greater resources available for exploration efforts. This excess of resources may cause ‘slack search’. This may also explain why firms that have both high exploitation as well as high exploration efforts generally perform better on the short term compared to non-ambidextrous firms.

We also found that environmental dynamism significantly influences the effect of ambidexterity on short-term performance (i.e. ROA). Ambidextrous firms in low dynamic environments perform better on the short term compared to ambidextrous firms in high dynamic environments. This finding may be explained by the rapid obsolesce of products and services in high dynamic environments. In low dynamic environments, firms may have less pressure to make large investments in order to renew, which subsequently may improve the

(30)

29 short term. This is in line with Jansen et al (2006), who also stated that the environment may moderate the effect of ambidexterity on performance, although they tested the effect of exploitation and exploration separately.

In addition, we found that in the short term, ambidextrous firms in high munificent environments outperform ambidextrous firms in low munificent environments. In these high munificent industries, there is more growth potential, which leads to less intense competition between the rival firms (Porter, 1979). Industries often follow a natural life cycle; they start with the pioneers, going to the growth phase, which results in the mature phase and finally will end at a decline phase (Castrogiovanni, 2002). High munificent environments are related with the first part of the industry life cycle. In these high munificent environments, the competitors’ actions do not harm the focal firms performance (Derfus et al., 2008), which leads to a higher short-term performance. Subsequently, the firms in low munificent environments are related with the latter part of the life cycle. Here, the competitors’ actions negatively influence the focal firm performance, which leads to a less optimal short-term performance. In addition, firms in high munificent environments generally have higher resources available to make the simultaneous pursuit of exploitation and exploration successful. Firms that have less resources available may not be able to afford such a complex strategy (Raisch, & Birkinshaw, 2008).

Both firms in low and high munificent environments benefit by pursuing ambidexterity. Firms in high munificent environments benefit because they already prepare themselves for the oncoming decline phase. Firms in low munificent environments benefit by ambidexterity because they can explore and look for new industries with more growth potential.

(31)

30 5.2 The effect of ambidexterity on firms’ long-term performance

As the results described in ‘Chapter 4’ indicated, ambidextrous firms have a higher Tobin’s Q ratio compared to non-ambidextrous firms. Here too, exploitation and exploration positively reinforce each other. This means that the simultaneous pursuit of exploitation and exploration adds to each other’s value. However, we did not find a significant interaction effect between ambidexterity and growth.

Furthermore, we found out that the relation between ambidexterity and Growth is moderated by munificence. Ambidextrous firms in high munificent environments have a higher growth ratio compared to ambidextrous firms in low munificent environments. From a logical point of view this makes sense, because growth and munificence (i.e. the growth potential of the industry) are interrelated with each other. Also, the relation between Tobin’s Q and Growth is moderated by munificence. Ambidextrous firms in high munificent environments have a higher Tobin’s Q ratio compared to ambidextrous firms in low munificent environments.

5.3 Limitations and further research

These results are obtained by analyzing the data of U.S. manufacturing firms that contain 500 or more employees. These results may therefore not be applicable to firms that 1) are located outside of the U.S., 2) are non manufacturing firms, or 3) contain less than 500 employees. In addition, there are some limitations regarding the use of patents as a measurement of exploitation and exploration. According to Ahuja (2002), not all the inventions are patentable and it also might be possible that inventions are not patented for a strategic reason. The directions for future research are derived from these limitations. Future research may clarify whether these findings may be different in other industries. Also, alternativemeasurements of exploitation and exploration could be investigated. In this research we operationalized those two concepts by building on patent data. A major limitation of building on patent data is that

(32)

31 not all innovations are patented (Hall, Jaffe, & Trajtenberg, 2001). This limitation would be avoided when one operationalized exploitation and exploration by not building on patent data.

6. Conclusion

The aim of this research is to contribute to a more fine–grained view of the impact of ambidexterity on firm performance. The central question of this paper is whether ambidextrous firms perform better compared to non-ambidextrous firms, if we take into account both the short- and long-term performance of the firm.

We can state that ambidexterity (i.e. the interaction between depth and scope) positively influences the short-term performance (ROA) of the firm. Firms that pursue ambidexterity are thus more likely to achieve superior performance in the short term compared to non-ambidextrous firms. In this research, we also tested whether and how environmental factors (i.e. environmental dynamism and munificence) moderate the relationship between ambidexterity and performance. The effect of ambidexterity on the short-term performance (ROA) is moderated by environmental dynamism as well as munificence. Ambidextrous firms in low dynamic environments generally perform better on the short term compared to ambidextrous firms in high dynamic environments. This may be caused by the quick obsolescence of the products in high dynamic environments, which leads to the necessity to invest in innovation and novelty. Also, ambidextrous firms in high munificent environments perform better compared to ambidextrous firms in low munificent environments in the short term. In low munificent environments, the rival actions negatively influence the focal firm performance, which leads to an inferior short-term performance (Derfus et al., 2008).

(33)

32 Besides the effect of ambidexterity on the short-term performance of the firm, we can also state that ambidexterity positively effects the long-term performance of the firm (Tobin’s Q). This implies that, also on the long term, ambidextrous firms perform better compared to non-ambidextrous firms. However, ambidexterity does not significantly influence the Growth of the firm, which is a long-term performance indicator as well. The effect of ambidexterity on the firm’s long-term performance (Tobin’s Q as well as Growth) is moderated by munificence. This relation goes the same way as with the short-term performance effects; ambidextrous firms in high munificent environments generally perform better in the long term compared to ambidextrous firms in low munificent environments.

We can thus conclude that, in general, ambidextrous firms benefit by pursuing exploitation and exploration simultaneously instead of pursuing one at the expense of the other.

(34)

33 References

Agarwal, R., & Helfat, C. E. (2009). Strategic renewal of organizations.Organization science, 20(2), 281-293.

Ahuja, G. (2000). Collaboration networks, structural holes, and innovation: A longitudinal study. Administrative science quarterly, 45(3), 425-455.

Atuahene-Gima, K. (2005). Resolving the capability—rigidity paradox in new product innovation. Journal of marketing, 69(4), 61-83.

Birkinshaw, J., & Gibson, C. (2004). Building ambidexterity into an organization. MIT Sloan management review, 45(4), 47.

Boyd, B. K. (1995). CEO duality and firm performance: A contingency model.Strategic Management Journal, 16(4), 301-312.

Burgelman, R. A. (1991). Intraorganizational ecology of strategy making and organizational adaptation: Theory and field research. Organization science,2(3), 239-262.

Cao, Q., Gedajlovic, E., & Zhang, H. (2009). Unpacking organizational ambidexterity: Dimensions, contingencies, and synergistic effects. Organization Science, 20(4), 781-796. Castrogiovanni, G. J. (2002). Organization task environments: Have they changed

fundamentally over time?. Journal of Management, 28(2), 129-150.

Chen, W. R. (2008). Determinants of firms' backward-and forward-looking R&D search behavior. Organization Science, 19(4), 609-622.

Cohen, J., Cohen, P., West, S. G. & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences.

Combs, J. G., Crook, T. R., & Shook, C. L. (2005). The dimensionality of organizational performance and its implications for strategic management research. Research methodology in strategy and management, 2(05), 259-286.

Derfus, P. J., Maggitti, P. G., Grimm, C. M., & Smith, K. G. (2008). The Red Queen effect: Competitive actions and firm performance. Academy of Management Journal, 51(1), 61-80. Dess, G. G., & Beard, D. W. (1984). Dimensions of organizational task

environments. Administrative science quarterly, 52-73.

Greve, H. R. (2003). A behavioral theory of R&D expenditures and innovations: Evidence from shipbuilding. Academy of Management Journal, 46(6), 685-702.

(35)

34 Gupta, A. K., Smith, K. G., & Shalley, C. E. (2006). The interplay between exploration and exploitation. Academy of management journal, 49(4), 693-706.

Hall, B. H., Jaffe, A. B., & Trajtenberg, M. (2001). The NBER patent citation data file: Lessons, insights and methodological tools (No. w8498). National Bureau of Economic Research.

He, Z. L., & Wong, P. K. (2004). Exploration vs. exploitation: An empirical test of the ambidexterity hypothesis. Organization science, 15(4), 481-494.

Jansen, J. J., Van Den Bosch, F. A., & Volberda, H. W. (2006). Exploratory innovation, exploitative innovation, and performance: Effects of organizational antecedents and environmental moderators. Management science, 52(11), 1661-1674.

Junni, P., Sarala, R. M., Taras, V., & Tarba, S. Y. (2013). Organizational ambidexterity and performance: A meta-analysis. The Academy of Management Perspectives, 27(4), 299-312. Katila, R., & Ahuja, G. (2002). Something old, something new: A longitudinal study of search behavior and new product introduction. Academy of management journal, 45(6), 1183-1194.

Kirby, J. (2005). Toward a theory of high performance. Harvard Business Review, 83(7), 30-39.

Lavie, D., Stettner, U., & Tushman, M. L. (2010). Exploration and exploitation within and across organizations. The Academy of Management Annals, 4(1), 109-155.

Levinthal, D. A., & March, J. G. (1993). The myopia of learning. Strategic management journal, 14(S2), 95-112.

MacKinnon, D. P. (2011). Integrating mediators and moderators in research design. Research on Social Work Practice, 21(6), 675-681.

March, J. G. (1991). Exploration and exploitation in organizational learning. Organization science, 2(1), 71-87.

Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Porter, M. E. (1979). How competitive forces shape strategy.

Raisch, S., & Birkinshaw, J. (2008). Organizational ambidexterity: Antecedents, outcomes, and moderators. Journal of management.

(36)

35 Romanelli, E., & Tushman, M. L. (1994). Organizational transformation as punctuated

equilibrium: An empirical test. Academy of Management journal,37(5), 1141-1166. Rowe, W. G., & Morrow, J. L. (1999). A note on the dimensionality of the firm financial performance construct using accounting, market, and subjective measures. Canadian Journal of Administrative Sciences/Revue Canadienne des Sciences de l'Administration, 16(1), 58-71. Simsek, Z. (2009). Organizational ambidexterity: Towards a multilevel

understanding. Journal of Management Studies,46(4), 597-624.

Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic management journal, 509-533.

Tushman, M., Smith, W., Wood, R., Westerman, G., & O’Reilly, C. (2003).Innovation streams and ambidextrous organizational designs: On building dynamic capabilities. Division of Research, Harvard Business School.

Tushman, M. L., & O'Reilly, C. A. (1996). The ambidextrous organizations: Managing evolutionary and revolutionary change. California management review, 38(4), 8-30. Tushman, M. L., & O'Reilly, C. A. (1997). Ambidextrous organizations: Managing

evolutionary and revolutionary change. Quality Control and Applied Statistics, 42, 215-218. Uotila, J., Maula, M., Keil, T., & Zahra, S. A. (2009). Exploration, exploitation, and financial performance: analysis of S&P 500 corporations. Strategic Management Journal, 30(2), 221-231.

Venkatraman, N., Lee, C. H., & Iyer, B. (2007). Strategic ambidexterity and sales growth: A longitudinal test in the software sector. In Unpublished Manuscript (earlier version presented at the Academy of Management Meetings, 2005).

Wang, H., & Li, J. (2008). Untangling the effects of overexploration and overexploitation on organizational performance: The moderating role of environmental dynamism. Journal of Management.

Wernerfelt, B., & Montgomery, C. A. (1988). Tobin's q and the importance of focus in firm performance. The American Economic Review, 246-250.

Zahra, S. A., & George, G. (2002). Absorptive capacity: A review, reconceptualization, and extension. Academy of management review, 27(2), 185-203.

Referenties

GERELATEERDE DOCUMENTEN

Het management van produktieprocessen kan men onderscheiden naar de beheersing van het primaire proces en de beheersing van een aantal ondersteunende processen.

Bester and Stanz (2007) noticing the SANDF’s growing role in peacekeeping operations raised a significant question, regarding the extent to which South African

In this study, we present and evaluate a robotically actuated delivery sheath (RADS) capable of autonomously and accurately compensating for beating heart motions by using

Het is opvallend dat Lamport zijn zeer theoretische werk – al zijn algoritmes zijn wiskundig beschreven en correct bewezen, al dan niet in TLA – altijd heeft uitgevoerd binnen

Moreover, the in vivo Aβ peptide pool is highly dynamic containing different Aβ peptides that interact and influence each other’s aggregation and toxic behaviour.. These Aβ

Effect of silent control (light grey, N ¼ 23) and short-term boat noise exposure (dark grey, N ¼ 17) playback on (a, d) the change in the time gobies were frozen before and

In model B, the added dummy variable for high levels of retention is positive and significant, meaning that retention rate has a significant positive influence on

Literature research facilitated the identification of three key concepts: namely, the characteristics of a family firm, individual long-term orientation and management control