‘‘Decoupling of R&D and Empty Green Patents as Response to Conflicting demands; and Increased Legitimacy Following This’’

(1)

‘‘Decoupling of R&D and Empty Green Patents as Response to

Conflicting demands; and Increased Legitimacy Following This’’

An Empirical Analysis of the Global Energy Sector for the years 2015-2017

MASTER THESIS

Business Administration:

MSc. Organizational & Management Control

MSc. Strategic Innovation Management

June 25

th

_{, 2018}

L.A. WOLF

S2492628

University of Groningen

Faculty of Economics and Business

(2)

ABSTRACT

Nowadays, firms are subject to two conflicting pressures. On the one hand, they face the pressure to attain short-term earnings, and, on the other, they are pushed to increase their environmental performance. This study suggests that executives decouple substantive and symbolic actions to comply to both pressures simultaneously. First, it is expected that they cut R&D, the internally-oriented, substantive part of innovation activities to respond to earnings pressure. Second, executives are expected to increase the level of green patents, particularly empty green patents, to comply to the pressure for environmental performance as well as to hide R&D cuts. The green patents represent the externally-oriented part of innovation activities and can therefore function as a signal for environmental performance and innovation. To study the effect of these pressures on decoupling, and ultimately legitimacy, 249 publicly-listed firms from the global energy sector are analysed for the years 2015-2017. The findings show that executives respond to earnings pressure and pressure for environmental performance with R&D cuts and empty green patents respectively. However, decoupling is mainly caused by the pressure for environmental performance. Further, decoupling has a positive effect on legitimacy in the eyes of the firm’s stakeholders and shareholders. Last, CEO/chair separation and a medium-sized board can limit the likelihood of decoupling. These findings demonstrate interesting avenues for future research and suggestions for policy makers. For instance, this research could be applied to other sectors or operational activities. Moreover, it is important for policymakers to develop measures that steer executives towards long-term value creation, so decoupling is less likely.

Keywords:

Conflicting demands, pressure for environmental performance, earnings pressure, empty green

patents, R&D cuts, decoupling of innovation activities, legitimacy, board composition, energy sector.

Acknowledgement:

I wish to thank dr. Joanna Gusc and prof. dr. Jordi Surroca for all their guidance and

(3)

1. INTRODUCTION

Firms are subject to multiple and sometimes conflicting pressures. On the one hand, firms are increasingly pressured to act in ways to enhance their environmental performance (Bansal & Roth, 2000; Bansal 2002, 2005; Starik & Kanashiro, 2013). This pressure originates from the extensive use of natural resources together with rapid economic growth, which have magnified environmental burdens and put a rise to major concerns by society (Berrone & Gomez-Mejia, 2009; Panwar, Kaushik & Kothari, 2011; Qi, Shen, Zeng & Jorge, 2010). On the other hand, firms face short-term earnings pressure since investors consider reported earnings as the most valued indicators to assess the attractiveness of a firm’s stock (Graham, Harvey, & Rajgopal, 2005; Kury, 2007; Lev, 1989; Rosenfield, 2000). At the same time, executives have incentives to achieve short-term earnings because their compensation is dependent upon earnings (Christensen, Kaufman, & Shih, 2010). Also, they feel pressured to attain short-term earnings targets to avoid being fired for insufficient performance (Roe, 2013). Thus, short-term earnings results are crucial if executives want to keep their job.

While earnings pressure sets off executives’ fixation on the short-term (Osma, 2008), the pressure for environmental performance urges executives to focus on the long-term. As a consequence, executives face a temporal tension when faced with both pressures contemporaneously. But how to deal with these conflicting pressures? Institutional theorists argue that decoupling–the separation of substantive and symbolic actions– enables firms to comply to diverging demands (Meyer & Rowan, 1977; Westphal & Zajac, 2001). So, executives can deal with the conflicting pressures by decoupling the firm’s activities in one important area of its operations: innovation. Hence, they can separate the internally-oriented, substantive part of innovation activities (like research and development (R&D) expenditures) from the externally-oriented, symbolic part of innovation activities (such as patents). Thus, to respond to earnings pressure, executives can cut R&D expenditures–the least visible part of innovation activities–to artificially boost short-term profits (Dechow & Skinner, 2000; Perry & Grinaker, 1994; Roychowdhury, 2006). However, at the same time, executives can increase green patents– the part of innovation activities that is highly visible to external stakeholders–to appear to keep constant the level of R&D and to deal with the pressure for environmental performance (Bansal & Gao, 2006; Dangelico & Pujari, 2010; Margolis & Kammen, 1999). Green patents are particularly visible since external stakeholders are pressuring the firm to enhance its environmental performance and are therefore closely observing the firm. In addition, patents would be effective signals for innovation activity since they are freely available to external observers (Graham, Mreges, Samuelson, & Sichelman, 2009). Ergo, innovation can play a double role and can be used as decoupling device to conform to both pressures simultaneously.

(6)

Elkhaldi, Atri, & Souid, 2011). Because prior studies found that executives were more likely to decouple when they were more powerful (Zajac & Westphal, 1995; Westphal & Zajac, 1998), it is expected that the board could play an important role in reducing executives’ opportunity to decouple innovation activities.

Although R&D cuts spurs short-term earnings, shareholders are likely to disagree since these cuts negatively affect a firm’s long-term profitability, and thereby shareholder value (Christensen et al., 2010; Guidara & Boujelbene, 2014). But, as corporate social responsibility (CSR) can be used as a shield for earnings management (Martínez-Ferrero, Banerjee, & Garcia, 2016), green patents can presumably function as a shield for R&D cuts. In this sense, green patents may enable executives to give an appearance of a constant level of innovation, and thus long-term financial success (Bansal, 2005; Kim & Lyon, 2015; Starik & Kanashiro, 2013). Accordingly, green patents can signal long-term shareholder wealth and may therefore lead to an increase in shareholders’ legitimacy. In addition, green patents can be used to signal environmental performance to the firm’s external stakeholders. Because green patents are highly visible, and the firm’s stakeholders increasingly attach value to the firm’s environmental performance (Lee, 2011), they can be quickly recognised. So, green patents can be used as signalling devices for environmental performance, and, in turn, can also positively affect stakeholders’ legitimacy.

So, in this study I propose actions taken by executives as an attempt to synchronously comply to conflicting pressures, whereby it will concentrate on how this affects legitimacy. More specifically, it will research if the tension between earnings pressure and environmental performance pressure leads to the decoupling of innovation activities. Therewith, executives may reduce R&D expenditures to achieve short-term earnings, whilst they simultaneously increase green patents. First, to comply with the pressure for environmental performance, and second, to mask their R&D cuts. Further, this study will look at how the firm’s board composition impacts this relationship. In this respect, the research questions central to this study are the following:

1. What is the influence of the pressure for environmental performance on green patents?

2. What is the influence of the earnings pressure on R&D cuts?

3. What is the influence of the pressure for environmental performance and earnings pressure on the

decoupling of innovation activities?

4. What is the impact of the firm’s board composition on the effect of the pressure for environmental

performance and earnings pressure on the decoupling of innovation activities?

5. How does the decoupling of innovation activities affect the firm’s legitimacy in the eyes of their

(7)

To examine the above questions, this study researched a sample of 249 firms from the global energy sector. This sector was chosen because it is subject to mounting environmental performance pressure. Energy firms experience much scrutiny from external stakeholders since they are portrayed as the main perpetrators of global warming (Ekwurzel et al., 2017). For instance, ExxonMobil, Shell, BP, and Chevron are identified as among the highest emitting firms since 1988 (The Guardian, 2017). On top of that, energy firms are vulnerable to earnings pressure resulting from market risks and operational risks caused by for example fluctuations in oil prices or the firm’s drilling success (Malmquist, 1990; Pincus & Rajgopal, 2002). Even though energy firms are prone to environmental pressure and earnings pressure, they still manage to be one of the most successful, making this sector particularly relevant to study.

In my empirical analyses, I made use of a unique international dataset for which I have collected data from different sources for the years 2015-2017 (Orbis, Thomson Reuters, Forbes, Environmental Performance Index, and World Justice Project). This dataset includes information about environmental performance pressure, analyst earnings forecasts, actual earnings, R&D expenditures, green patents, stock market values, the Forbes Global2000 ranking, board composition, SIC codes, country of origin, total assets, revenue, debt, and slack.

The results of this study show that pressure for environmental performance and earnings pressure have a positive effect on green patents and R&D cuts respectively. Additionally, the pressure for environmental performance positively influences decoupling. However, no effect was found for earnings pressure on decoupling. Regarding board composition, a negative moderation effect was found for CEO/chair separation and a medium-sized board. Last, decoupling leads to higher levels of legitimacy in the eyes of the firm’s stakeholders as well as shareholders, implying that it may be enough if firms solely create an illusion of complying to different demands to increase legitimacy.

(8)

2. THEORETICAL FRAMEWORK

In this section, I describe the theories used to frame this study, namely: institutional theory, including the notion of decoupling, and agency theory.

2.1 Theory

2.1.1 Institutional Theory

According to institutional theory, firms are embedded in institutional environments. Actors in these environments have certain expectations of firms, meaning that there are normative pressures–certain values, norms and rules–that firms are expected to act upon (Boxenbaum & Jonsson, 2008). Firms need to abide by these expectations to maintain legitimacy (Scott, 2005). Legitimacy can be defined as ‘‘a generalised perception

or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions’’ (Suchman, 1995: p. 574). Firms need legitimacy

in order to survive and prosper (Boxenbaum & Jonsson, 2008; Meyer & Rowan, 1997). Complementary to normative pressures, firms face coercive pressures resulting from laws. In the case of coercive pressures, firms need to conform to them since failure to do so is illegal and illegitimate (Edelman, 1992). To remain legitimate, firms are forced to engage in practices that are valued by their institutional environment.

In the last few years, pressure for environmental performance has risen. For instance, firms are increasingly subject to inspections for environmental wrongdoing, investors discount stock prices of polluting firms, governments enforce more and more environmental policies, and customers attach value to the environmental quality of a firm's products. In this vein, firms are ever more pressured by regulators and legislators as well as other stakeholders to enhance their environmental performance (Campbell, 2007; Delmas & Toffel, 2008). Idem, firms are expected to increasingly invest in green innovation when their institutional environment pressures them to do so (Berrone, Fosfuri, Gelabert, & Gomez-Meija, 2013; Brunermeier & Cohen, 2003; Campbell, 2007, Lee 2011).

2.1.1.1 Decoupling

(9)

Decoupling means that executives separate two types of activities: substantive and symbolic (Ashforth & Gibbs, 1990; Meyer & Rowan, 1977; Oliver, 1991). Substantive actions usually present inward-looking practices that involve real actions aimed for the development of organisational capabilities (e.g. Fligstein, 1985; Kamens, 1997; Meyer & Rowan, 1997). These actions may sometimes constrain the firm’s internal flexibility since they often ask for significant changes or long-term investments (Eccles, Ioannou, & Serafeim, 2014). In contrast, symbolic actions are externally-oriented because they include public and highly visible activities. These actions are generally undertaken to seek public endorsement of the firm and its practices (Hawn & Ioannou, 2015). Executives' motives to decouple often stem from their desire to keep doing business as usual, while appearing to comply to diverging institutional expectations. So, executives favour symbolic rather than substantive changes (Meyer & Rowan, 1977; Suchman, 1995). With regard to innovation activities, symbolic/externally-oriented actions would include patents because they are publicly visible. Consequently, they can be used to communicate the firm’s innovative activities to its external environment. Green patents in particular can be easily observed by external stakeholders since a firm’s environmental performance is subject to increasing levels of public scrutiny these days (Bilbao-Osorio et al., 2012; Johnstone, Hascic & Ostertag, 2008). R&D expenditures, on the other hand, would refer to the firm’s substantive activities because they encompass internally-oriented, long-term investments served to foster the firm’s innovative capability (Bereskin et al., 2018; Christensen et al., 2010; Fiss & Zajac, 2006).

At the present-time, the increased concerns about sustainability and environmentally responsible behaviour exceed firms’ accountability beyond financial accountability. Nevertheless, short-term profit maximisation still stays an important goal (Yip, Stade, & Cahan, 2011). This implies that executives face a tension between short-term earnings and environmental performance. As a result, executives may engage in decoupling to seek compliance to these diverging demands (Westphal & Zajac, 2001). So, to respond to earnings pressure as well as to the pressure for environmental performance, executives may decouple innovation activities. Hence, they may cut R&D expenditures–the invisible part–to boost short-term earnings, whereas they may increase green patents–the highly visible part–to respond to the pressure for environmental performance. Hereby, executives respond to external expectations by merely symbolic action in order to seek ceremonial compliance without making any positive substantive change (Ashforth & Gibss, 1990). In fact, they are even decreasing the substantive, internally-oriented innovation activity, which could possibly be hidden behind the increase in patents. That green patents can serve as a mask for R&D cuts is based on the notion of Martínez-Ferrero et al. (2016) which states that CSR can be used as a shield for EM. In this regard, green patents can be the symbolic signal to the firm’s stakeholders and shareholders that it cares for the long-term and the environment (Siano, Vollero, Conte, & Amabile, 2017), while it is actually cutting R&D, and therewith future firm value.

(10)

comply with different demands in order to assure legitimacy in the eyes of for example the firm’s shareholders, the general public, financial community, and regulators (Meyer & Rowan, 1977; Tilscik, 2010; Westphal & Zajac, 2001).

2.1.2 REM from the Agency Theory Perspective

Prior literature has established a relation between agency theory and earnings management (EM) (e.g. Davidson III, Jiraporn, Kim, & Nemec, 2004). Agency theory argues that the separation of ownership and control may lead to conflicts of interests between executives (agents) that run the firm and shareholders (principals) that own the firm (Jensen & Meckling, 1976). As opposed to shareholders, executives have a short-term focus. Consequently, this may lead to opportunistic behaviour because executives often serve their short-term interest, even though this may sacrifice shareholders’ long-term interest (Martínez-Ferrero et al., 2016; Merchant & Van der Stede, 2012). Hence, EM can be related to agency theory because it demonstrates a conflict of interests between executives and shareholders between short-term and long-term earnings.

EM can be used to artificially boost short-term earnings (Burgstahler & Dichev, 1997; Dechow & Skinner, 2000). According to Schipper (1989), EM can be divided into two types: accounting (AEM) and real earnings management (REM). With the former, executives take advantage of flexibility in accounting standards to conceal the firm’s true economic performance, without violating accounting rules (Dechow, Sloan, & Sweeney, 1995). With the latter, executives take actions that deviate from normal business practices with the primary objective of meeting certain earnings thresholds (Roychowdhury, 2006). Executives have incentives to practice EM because executives’ turnover is associated with weaker short-term earnings performance (Roe, 2013). So, EM is driven by executives’ desire to remain qualified for their position. In addition, it is tempting for executives to practice EM in order to secure their own payment, which is often based on short-term earnings performance (Dye, 1988; Guidara & Boujelbene, 2014; Ola & Abad, 2014; Trueman & Titman, 1988).

Unlike AEM, REM threatens the firm’s future value creation (Bereskin, Hsu & Rotenberg, 2018). Because executives are only in the firm for about seven years (Kaplan & Minton, 2012), they probably do not have to suffer the long-term consequences of REM, whereas they can still enjoy the advantages of short-term earnings performance (Osma, 2008). As a result, it may be more attractive for executives to engage in REM instead of AEM.

(11)

relatively short period of time. This means that good earnings performance can happen on their watch, while poor earnings results are pushed into the future, thus beyond their tenure (Roe, 2013). In addition, R&D cuts can be used for REM since there is a high degree of information asymmetry associated with R&D (Hall, 2002; Holmstrom, 1989). This implies that R&D cuts–for the purpose of spurring short-term earnings–are not easily visible nor detected by others. Further, as executives are the most important decision-makers within firms, they can use their power and discretion to make decisions regarding R&D activities. Thus, when there is no appropriate corporate governance mechanism in place, executives are not effectively monitored, meaning that they can use their power to cut R&D (Guidara & Boujelbene, 2015; Roe, 2013). So, executives’ power, desire to remain in their positions, and short-term vision are likely to result in R&D cuts, even though these cuts may be detrimental to long-term shareholder value.

2.2 Hypotheses Development

2.2.1 Empty Green Patents and Pressure for Environmental Performance

Patents are intellectual property rights granted to an inventor in exchange for public disclosure of an invention. They refer to legally granted rights that exclude others from making, using, selling, or importing an invention, for a limited time, within a certain country (James, Leiblein, & Lu, 2013). Green patents are granted for green inventions, which are proxies for green innovation (Kemp, 2008). Green innovation is ‘‘the production,

assimilation or exploitation of a product, production process, service, management or business method that is new to the firm and incorporates energy conservation, prevention of pollution, recycling of waste, green product design, and/or corporate environmental management’’ (e.g. Aguilera-Caracuel & Ortiz-de-Mandojana, 2013;

Lai, Wen, & Chen, 2003; Kemp & Pearson, 2007). The overall purpose of green innovations is the reduction of harmful environmental impact (Urbaniec, 2015). So, green innovations are inventions of new designs and the creation of novel products, processes, services, management or business methods with improved environmental standards (Berrone et al., 2013).

Based on insights from institutional theory, pressure for greening impacts firms’ attempts to invest in green innovation. In a similar fashion, Campbell (2007) and Lee (2011) argue that firms increasingly attach value to their environmental performance when their institutional environment deems this important. As green patents are highly visible to external stakeholders (Graham et al., 2009), it is expected that firms increase green patents to respond to the pressure for environmental performance.

(12)

Jaffe, Newell, & Stavins, 2005; Rennings, 2000). Because executives usually have a short-term tenure, they do not profit from the long-term advantages of valuable green innovations, and may therefore not want to incur the short-term costs. (Davit, Hitt, & Gimeno, 2001). Hence, executives may increase the patents that represent less valuable innovations, namely the patents without citations (empty patents). According to Trajtenberg (1990), patents that were cited (more) cover inventions of greater social value. Moreover, a U.S.-based study found that a single citation increased the economic value of a patent with more than $1 million (Harhoff, Narin, Scherer, & Vopel, 1991). Based on this, empty patents refer to green innovations that are of lesser economic and social value.

Following the above and the argument of Oliver (1991) that says that: ‘‘the appearance rather than the fact of

conformity is often presumed to be sufficient for the attainment of legitimacy’’ (p.155), an increase in empty

green patents may be enough to conform to institutional pressures. Therefore, the pressure for environmental performance can give rise to solely symbolic responses, served to change the firm’s reputation instead of its behaviour (Deegan & Gordon, 1996). In this sense, empty patents can be an effective way to signal compliance to the pressure for environmental performance. This leads to the following hypothesis:

H1. Pressure for environmental performance positively affects the level of empty green patents.

2.2.2 R&D Cuts and Earnings Pressure

Researchers have shown that executives engage in REM to achieve short-term earnings targets (Baber, Fairfield, & Haggard, 1991; Bange and De Bondt, 1998; Bushee, 1998; Cheng, 2004; Graham et al., 2005; Osma, 2008; Osma & Young, 2009; Perry and Grinaker, 1994), to outdo analysts’ earnings forecasts (Bhojraj, Hribar, Picconi, & McInnis, 2009), and to seize earnings-based compensation (Bange and De Bondt, 1998; Harter & Harikumar, 2004). So, executives have incentives to artificially boost earnings via REM (Dechow & Skinner, 2000; Osma & Young, 2009). Roychowdhury (2006) defines REM as ‘‘management actions that deviate from

normal business practices, undertaken with the primary objective of meeting certain earnings thresholds’’ (p.

336).

As argued in the theory section, REM can be practiced by cutting the level of R&D expenditures. Several researchers found a link between earnings pressure and R&D cuts. For instance, Demirag (1998) says that only ten percent of a firms R&D expenditures are insensitive to lagged earnings performance. Furthermore, the study of Graham et al. (2005) shows that executives see R&D investments as expenses that they can cut when earnings targets are not met. Also, Osma (2008) states that executives may choose to cut R&D when faced with earnings pressure. For these reasons, the following hypothesis is formulated:

(13)

2.2.3 Decoupling of Innovation Activities

Sougiannis (1994) estimated that an increase in R&D by one dollar leads to an increase in profits of two dollar over a period of seven years. Consequently, R&D is critical for a firm’s long-term performance (Lee & O’Neill, 2003). Thus, if executives cut R&D to boost earnings, they decrease chances of long-term performance and innovative growth (Bereskin et al., 2018; Christensen et al., 2010). Because R&D cuts destroy long-term firm value, (Bhojraj, Hribar, Picconi, & McInnis, 2009), shareholders are likely to disagree with R&D cuts. This would imply that executives may be motivated to mask these R&D cuts in order to keep shareholders’ support and remain in their positions.

Several scholars argue that CSR can be used to cover up EM. For example, Martínez-Ferrero et al. (2016) argue that executives use CSR as a shield for EM. Further, McWilliams, Siegel and Wright (2006) state that CSR can be used to advance executives career or to disguise deviant behaviour. In accordance, Hemmingway and Maclagan (2004) say that executives may engage in CSR to ensure their reputation, especially in the presence of EM. In a similar vein, it can be argued that executives use empty green patents to hide R&D cuts.

In addition, empty green patents can be used as an attempt to signal conformity to stakeholders’ expectations regarding environmental performance. These patents can be especially effective since they are expected to be quickly observed by the firm’s stakeholders when pressure for environmental performance exists.

So, executives can use R&D cuts to boost short-term earnings. Simultaneously, they can increase empty green patents to respond to the pressure for environmental performance and to hide these R&D cuts. Hereby, the increase in empty green patents is a symbolic and externally-oriented action that serves as a signal for long-term firm and environmental performance. Hence, they may deflect attention from less admirable activities (Morris & King, 2010), namely the R&D cuts.

Bearing the above in mind, it is expected that executives decouple innovation activities when faced with earnings pressure on the one hand and pressure for environmental performance on the other. This leads to the following hypothesis:

(14)

2.2.4 Board Composition

Corporate governance determines the relationship between the firm’s owners (shareholders) and executives, which is central to the concept of agency (Jensen & Meckling, 1976). The firm’s board of directors is critical to corporate governance and central to decision-making (Malmi & Brown, 2008). Boards should monitor executives and should ensure that a firm’s assets are managed efficiently and according to the interests of the providers of finance (Kouki et al, 2011; Osma, 2008; Osma & Noguer, 2007; Shleifer & Vishny, 1997). However, different board characteristics can strengthen or weaken the board’s monitoring ability. Therefore, this study suggests that board composition influences executives’ opportunity to decouple innovation activities.

2.2.4.1 Board size

Board size1_{is an important factor of board efficiency (Kouki, Elkhaldi, Atri, & Souid, 2011). A large board size}

means diversity of opinions and a broader range of experience, thus better control. However, the amount of coordination and communication problems increases with board size (Yermack, 1996). So, board should neither be too big nor too small, whereby the optimal size to reduce agency problems is expected to be between five and nine members (Campos, Newell & Wilson, 2002). Accordingly, I expect that a board of five to nine members (medium-sized board) is negatively related to the decoupling of innovation activities.

2.2.4.2 Independent directors

Agency theory literature suggests that boards that have a higher share of independent directors2_{have superior}

governance (Core, Holthausen, & Larcker, 1999; Dahyaa & McConnell, 2005). In accordance, Osma (2008) states that monitoring and control by the board is enhanced by incorporating independent directors. Consequently, the likelihood of opportunistic behaviour is reduced. Furthermore, several scholars found that higher board independence is associated with a lower degree of EM and R&D cuts (e.g. Eng & Mak, 2003; Gul & Leung, 2004; Osma, 2008). Therefore, I expect that that the higher the share of independent directors, the less decoupling of innovation activities.

2.2.4.3 CEO/chair separation

When the CEO (Chief Executive Officer) is also the chairman of the board, the CEO has more power. In addition, this allows the CEO to select which information is available to other board members. As a result, monitoring effectiveness decreases (Jensen, 1993). Thus, when the CEO is not the chairman of the board (CEO/chair separation), monitoring will be higher and opportunistic behaviour is less likely. For these reasons, this study suggests that CEO/chair separation is negatively related to the decoupling of innovation activities.

1_{It refers to the number of administrators.}

(15)

Based on the above, I expect an appropriate board composition, which is medium-sized, has a higher share of independent directors, and has CEO/chair separation, to be effective in reducing opportunistic behaviour, and therefore the decoupling of innovation activities. In this vein, the following hypothesis is proposed:

H4. An appropriate board composition (medium-sized, independent, and separation of the CEO and the chair) negatively moderates the effect of pressure for environmental performance and earnings pressure on the decoupling of innovation activities.

2.2.5 Legitimacy

As discussed in the theory section, legitimacy is ‘‘a generalised perception or assumption that the actions of

an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions’’ (Suchman, 1995: p. 574).

Efforts for firms to enhance environmental performance are expected to impact legitimacy (Dhavele & Sarkis, 2015). As green patents can signal environmental performance, they supposedly increase the firm’s legitimacy in the eyes of its institutional environment. Also, because the appearance of conformity is often enough for sustaining legitimacy (Oliver, 1991), and ‘‘legitimacy flows from symbols’’ (Roa, 1994: p. 30), empty green patents alone can be sufficient for a firm to gain legitimacy in the eyes of the firm’s stakeholders.

Further, since firms with greater patent performance have greater market valuation (Lanjouw & Shankerman, 2004; Lerner, 1994), it can be argued that shareholders attach value to the firm’s patenting activity. Following this, it is expected that firms are able to get more support from their shareholders when they increase the level of green patents. Moreover, the achievement of earnings targets resulting from R&D cuts enable executives to comply to earnings pressure imposed by financial markets. Also, the increase in green patents allows executives to hide these cuts as well as to signal innovation and, therewith, long-term performance to the firm’s shareholders. As a consequence, decoupling is expected to increase legitimacy in the eyes of the firm’s shareholders.

Taking the above into account, the decoupling of innovation activities may lead to increased legitimacy in the eyes of the firm’s stakeholders and shareholders. Therefore, the following hypotheses are formulated:

(16)

2.3 Conceptual Model

The conceptual model presented in Figure 1 provides an overview of the variables and hypotheses tested in this study. All hypothesised relationships are tested at the firm-level.

Figure 1. Conceptual Model

3. METHODOLOGY

(17)

3.1 Sample and Data Collection

This study is based on the global energy sector3_{for the years 2015-2017. Initially, there were 3187 firms}

operating in energy industries to be found in Orbis. However, I chose to only focus on publicly-traded firms (universities, private firms, and public national research institutions were excluded from the sample because they are presumably not faced with earnings targets imposed by analysts). After deleting firms that were not publicly-traded and did not have all the necessary financial data available, 997 firms were left. When looking for board composition data in Thomson Reuters, there were 306 firms remaining. Finally, after collecting data on the pressure for environmental performance and earnings pressure, again firms were dropped resulting in a final sample of 249 firms.

Financial and operational data was obtained from Orbis (Bureau van Dijk). Further, patent data from the European Patent Office’s (EPO) PATSTAT dataset was derived from Orbis. Moreover, data on board composition was collected from the Thomson Reuters database, which is validated in prior literature (e.g. Hawn & Ioannou, 2016). In addition, data on the pressure for environmental performance was retrieved from the Environmental Performance Index (EPI) website and the World Justice Project (WJP) website. Last, the website of Forbes was consulted to retrieve data on the ranking used to determine stakeholders’ legitimacy.

3.2 Dependent Variables

3.2.1 Empty Green Patents

Patent data is seen as the most important measure of innovation (Berrone et al., 2013; Frietsch & Grupp, 2006). To identify green patents, the Y02 Cooperative Patent Classification (CPC) is applied. This classification was developed by the EPO, the United Nations Environmental Program, and the International Centre on Trade and Sustainable Development (Aguilera-Caraacuel & Ortiz-de-Mandajana, 2013). Patent citations reflect the value of innovations. So, empty patents (without citations) would imply less valuable innovations since they generally have less economic and/or social impact (Trajtenberg 1990; Hall, Jaffe, & Trajtenberg, 2005). In sum, to

measure empty green patents (PATENTS) 4, this study looks at the increase in the number of granted, non-cited

Y02 patents in year t (2016 as base year).

3_{I defined the following industries to be part of the energy sector: mining of coal and lignite; extraction of crude petroleum}

and natural gas; support activities for petroleum and natural gas; manufacturing of coke and refined petroleum products; processing of nuclear fuel; electricity, gas, steam, and air conditioning supply; wholesale of solid, liquid, and gaseous fuels and related products. I based this on the types of energy sources used worldwide and searched in Orbis for the matching industries. I retrieved these industry names from Orbis based on NACE rev. 2.

4_{The names of the variables between parentheses are used in the tables in the results section as well as in the regression}

(18)

3.2.2 R&D Cuts

To identify R&D cuts (RD_CUT), the abnormal levels of R&D first needed to be calculated, which I based on Roychowdhury (2006). Equation 1 shows the regression that is used to calculate abnormal levels of R&D.

R&D_EXP is the total amount of R&D expenditures for firm i and year t, TA are the total assets for firm i and

year t-1, and REV is the operating revenue for firm i and year t-1. The regression is estimated for each industry

(two-digit SIC code) in the yeart. Industries with fewer than 15 firms are eliminated from the sample since the

residual cannot be calculated. !&#_%&'_(,* +,(,*-. =a0+ a. 3 1 +,(,*-.5 + a6 3 !%7_(,*-. +,(,*-.5 + e(.* 9&: Equation 1

The residual e R&D_{measures the level of abnormal R&D. A lower value}e R&D_{indicates a deeper opportunistic}

cut in firm i’s R&D expenses. Therefore, e R&D_{(the residual) will be multiplied by minus one. Hence, a higher}

value of RD_CUT demonstrates a higher R&D cut (Equation 2). !#_;<+_(,* = −1 > e_(,*9&:

Equation 2

3.2.3 Decoupling of Innovation Activities

A dummy variable is created to measure the decoupling of innovation activities (DECOUPLING), which equals 1 if R&D_CUT (equation 2) is above zero and the firm has also increased the number of empty green patents

in year t, and 0 otherwise.5

3.2.4 Legitimacy

3.2.4.1 Stakeholders’ Legitimacy

Legitimacy in the eyes of the firm’s stakeholders (SK_LEG) is measured using the Forbes Global2000 ranking

of 20176_{. The use of a ranking as a measure for legitimacy is validated in prior studies (e.g. McDonnel & King,}

2013). The values 1-95 were assigned to firms that were placed in the ranking, where the value 95 was assigned to the firm positioned highest in the ranking and 1 to the firm positioned lowest in the ranking. Firms that did

not make it to the Forbes ranking were given the value 0.7_{The ranking was chosen as a measure of legitimacy}

5_{This measurement is also used as independent variable for the effect of decoupling on both types of legitimacy.}

6_{The Forbes Global2000 is a ranking composed of equally-weighted measures of market value, revenue, profits, and assets.} 7_{Because the ranking consisted of 2000 positions and there can be a big difference in position between firms placed high}

(19)

since reputational rankings are an increasingly prominent part of the business environment of firms and they are permeating in almost all industries (Cleaver, 2003; Fombrun 1996). Furthermore, rankings are ‘‘certification

contests’’ that generate status orderings (Rao, 1994: p. 29), making this a suitable measure for legitimacy. 3.2.4.2 Shareholders’ Legitimacy

Legitimacy in the eyes of the firm’s shareholders (SH_LEG) is measured by looking at the market price per share in the year t+1, reflecting firm value (Díez-De-Castro & Peris-Ortiz, 2018).

3.3 Independent Variables

3.3.1 Pressure for Environmental Performance

Pressure for environmental performance (ENV_PRESS) is the principal component of two indexes (previously

standardised), whereby the Cronbach’s alpha coefficient was above 0.78_{. The first index used is the}

environmental performance index (EPI) per country9_{. The other index used is the Rule of Law Index}10_{. In this}

study, I have assumed that the stronger and better the rule of law and the greater the environmental performance in a given country, the more pressure there is for the firm to increase its environmental performance.

3.3.2 Earnings Pressure

Earnings pressure (EARN_PRESS) is measured by the difference between analysts’ earnings forecasts per share and actual earnings per share in year t-1. Analysts’ forecasts per share are calculated by the average of all analysts’ earnings estimates in the prior period (Kasznik and McNichols 2002).

3.4 Moderator and Control Variables

3.4.1 Board Composition

Board composition is measured by looking at board size, the share of independent directors, and CEO/chair separation. Medium-sized board (MS_BOARD) is a dummy variable which receives the value 1 when the board has between five and nine members, and 0 otherwise. The share of independent directors (IND) is measured by

the ranking also received the highest score. This categorisation was inspired by the measurement of McDonnel and King who scored firms based on their position in the ranking. After, assigning values to each firm, the measure was standardised.

8_{The Cronbach alpha refers to the statistical power of the summary.}

9_{EPI (2014) is a joint project between the Yale Centre for Environmental Law & Policy and the Centre for International}

Earth Science Information Network at Columbia University, in collaboration with the Samuel Family Foundation and the World Economic Forum.

10_{The Rule of Law Index (2015) is developed by the WJP and measures the rule of law based on experiences and}

(20)

the percentage of outsiders on the board. CEO/chair separation (CHAIR_SEP) is a dummy variable that equals 1 when the CEO is not the chairman of the board, and 0 otherwise.

3.4.2 Control Variables

In order to ensure that the dependent variable is explained by the independent variable(s), several control variables are included to control for their confounding effects. In this section, the measurement and the expected effects of the control variables are discussed.

3.4.2.1 Firm Size

Firm size (F_SIZE) is measured by the natural logarithm of total assets to reduce potential effects of extreme values (Mishina, Dykes, Block, & Pollock, 2010). Larger firms are followed by analysts more (Wiedman, 1996) which reduces opportunities for R&D cuts. Further, larger firms are expected to enjoy better reputation (Brammer & Pavelin, 2004), and thus legitimacy.

3.4.2.2 R&D Intensity

R&D intensity (RD_INT) is measured by dividing total R&D expenses by revenue. High R&D intensive firms are expected to have less R&D cuts (Osma & Young, 2009). Also, this measure controls for a firm’s critical mass to innovate.

3.4.2.3 Leverage

To control for firm leverage (LEV), the firm’s leverage ratio is used. To determine the leverage ratio, I divided the firm’s total liabilities and debt by its total assets. Leverage ratio is expected to increase the likelihood to cut R&D (Osma, 2008).

3.4.2.4 Slack

Available slack (SLACK), measured by the firm’s current ratio, is used as indicator of slack. Firms with more slack have less need to engage in risky actions (Cyert & March, 1963; Greve, 2003).

3.4.2.5 Growth

Growth (GROWTH) is measured by the percentage increase in revenue compared with the prior year (Dadbeh

& Mogharebi, 2013) A primary characteristic of earnings manipulators was high growth. This is because the

(21)

3.4.2.6 Industry

There is controlled for industry (INDUSTRY) type using four-digit Standard Industry Classification (SIC). Three categories of SIC codes are identified in the sample based on the study of Waddock & Graves (1997). Multiple dummies will be created whereby the firm receives the value 1 if it falls within the range of SIC codes shown below, and 0 otherwise.

Table 1. Industry Classifications

Industry SIC Frequency

Mining and construction 100-1999 224

Refining, rubber, and plastics 2891-3199 77

Industry, telephone, and utilities 4732-4991 88

Total 249

3.4.2.7 Region

Multiple dummies are created to control for regional effects. Table 2 demonstrates the regions (REGIONS) identified in the sample.

Table 2. Regions

Region Frequency Asia 36 Europe 81 North America 90 Oceania 25 South America 17 Total 249

3.5 Data Analysis

(22)

Different types of regressions are used to test the hypotheses in this study. For the regression equations, I refer to Appendix III. First, to test the effect of pressure for environmental performance on empty green patents (hypothesis 1), I use a negative binominal regression (NBR). An NBR is be used since the dependent variable (PATENTS) is a counting variable implying that it is non-negative (Gardner, Mulvey, & Shaw, 1995). Next, I use a linear regression (LR) to test the effect of earnings pressure on R&D cuts (hypothesis 2). A LR is chosen because it allows for estimating relationships between variables that are continuous (Tripepi et al., 2008). Hereafter, I test the effect of the pressure for environmental performance and earnings pressure on the decoupling of innovation activities with a probit regression (PR) (hypothesis 3). A PR is appropriate since the dependent variable is a dummy (binary) (Horowitz & Savin, 2001). For each of the regressions, I have first tested the model with the control variables only. Afterwards, I have included the independent variable(s) in the model.

To look for moderation effects, I have mean-centred the variables earnings pressure, pressure for environmental

performance, and share of independent directors (Aiken & West, 1991)11_{. I used these mean-centred variables}

to create interaction effects, whereby I multiplied both independent variables separately with the different variables of board composition: medium-sized board (dummy), share of independent directors, and CEO/chair separation (dummy). To test the moderation effect of board composition, I used PRs (hypothesis 4). I ran multiple PRs for each variable separately (medium-sized board, the share of independent directors, and CEO/chair separation). In the first PR, I included the direct effect. Hereafter, I included the direct effect and the moderation effect with the pressure for environmental performance. Then, I ran a PR on the direct effect and the moderation effect with earnings pressure. Last, I included the direct effect and the moderation effect with

both the pressure for environmental performance and earnings pressure (Dawson, 2014)12_.

In addition, the effect of decoupling on stakeholders’ legitimacy is tested with a TR (hypothesis 5).13_{Last, I use}

a LR to test the effect of decoupling on shareholders’ legitimacy (continuous variable) (hypothesis 6).

Because multicollinearity may harm the reliability of the estimators of the model parameters (Alin, 2010), I have checked the Variance Inflation Factors (VIF) of the variables. Literature argues that the VIF value should not exceed the threshold of ten (e.g. O’brien, 2007; Robinson & Schumacker, 2009). VIF scores were all below five, so I have no reasons to suspect issues of multicollinearity.

11_{The mean of the variable is subtracted from the original value, so the mean-centred variable has a mean of zero.} 12_{Dawson (2014) was followed to study moderation effects of board composition. It says that the moderator variables as}

well as the interaction term of the moderator with the independent variable needs to be included to test the moderation effect.

13_{For H5, I used a TR following Iman and Conover (1981) and Ballenberger, Lluis, von Mutius, Illi, and Schaub (2012).}

(23)

4. RESULTS

This section presents the results of this study. First, the descriptive statistics and correlations are discussed. Hereafter, the results of the regression analyses are presented.

4.1 Descriptive Statistics and Correlations

The descriptive statistics, including the mean, standard deviation, minimum and maximum value, and correlations are shown in Table 3. The sample consists of 249 firms and no high values of correlations are present (r<0.8). Decoupling is significantly positively correlated with stakeholders’ and shareholders’ legitimacy (r=0.260 and r=0.233, p<0.01), indicating that decoupling may lead to more legitimacy. Further, decoupling is significantly and positively correlated with R&D cuts and empty green patents, which should be the case since the dummy variable referring to decoupling is created out of those. Moreover, pressure for environmental performance is significantly positively correlated with decoupling (r=0.104, p<0.1), while earnings pressure is not correlated with decoupling (r=0.007, p>0.1). This indicates that the pressure for environmental performance may lead to more decoupling, whereas earnings pressure may not necessarily be associated with decoupling. Also, a significant positive correlation exists between earnings pressure and R&D cuts (r=0.278, p<0.01), pointing to more cuts when faced with higher earnings pressure. In addition, pressure for environmental performance is slightly significantly and positively correlated with earnings pressure, implying that firms may face more pressure for environmental performance when they face more earnings pressure, and vice versa. Last, none of the variables of board composition is correlated with decoupling.

4.2 Regression Results

(24)

Table 3. Correlation Matrix and Descriptive Statistics

Number of observations: 249; * p < 0.10, ** p < 0.05, *** p < 0.01; all variables are winsorised at 1-99 percent level.

(25)

Table 4. NBR on Empty Green Patents

Table 5. LR on R&D Cuts

Number of observations: 249; * p < 0.10, ** p < 0.05, *** p < 0.01; all variables are winsorised at 1-99 percent level. Industry and region effects have been included in the analysis.

Table 5 exhibits the results of hypothesis 2. Earnings pressure is positively and significantly related to R&D

cuts (Model 2: b=0.0356, p<0.05)14_{. Thus, support is found for H2. In addition, firm size significantly negatively}

impacts R&D cuts (p<0.01). For instance, in model 1, we can see that firms size lowers R&D cuts with 13.41%. In contrast, R&D intensity significantly positively influences R&D cuts (p<0.01). Leverage shows a positive relation with R&D cuts in model 1 (p<0.1). Further, growth has a positive significant relationship with R&D cuts (p<0.05). Slack has no effect.

Table 6 demonstrates the results of hypothesis 3 and 4. Hypothesis 3 is partially supported. The pressure for environmental performance has a positive and significant relation with decoupling (Model 1: b=1.5744, p<0.01) and earnings pressure is not related to decoupling (p>0.1). Also, CEO/chair separation negatively and

significantly moderates the effect of earnings pressure on decoupling (Model 13: b=-1.0092, p<0.05). A

medium-sized board negatively and significantly moderates the effect of earnings pressure on decoupling

(Model 5: b=-0.6086, p<0.1)15_{. No moderation effects were found for the share of independent directors. In}

addition, firm size positively and significantly influences decoupling (p<0.01, p<0.5). Further, leverage positively and slightly significantly affects decoupling in 3 out of 14 models (p<0.1), and slack positively and significantly influences decoupling in all models (p<0.05). No effects were found for the other controls.

14_{Because there is much ambiguity about the way of measuring R&D cuts and earnings pressure, an additional regression}

with a different measure of both variables is performed. Here, a dummy was created for R&D cut that equalled one if R&D was lower in year t than in year t-1, and zero otherwise. Also, to measure earnings pressure, a dummy was created that equalled one if earnings before taxes in (EBT) t-1 was higher than EBT in t-2, however, EBT plus R&D expenses t-1 was lower than EBT plus R&D in t-2. The correlation between R&D cut and earnings pressure was positive (r=0.1998, p<0.01). Additionally, earnings pressure showed a significant positive effect on R&D cuts (p<0.1).

15_{Also, since there is much ambiguity about the ideal board size, two extra tests have been performed to see if other}

measures of board size had an effect on decoupling. PRs with the number of board members as continuous variable and a dummy for a board of 8-10 members both showed no effect on decoupling.

VARIABLES Model 1 Model 2 VARIABLES Model 1 Model 2

Controls Controls & independent variable

(26)

Table 6. PR on Decoupling

Number of observations: 207; * p < 0.10, ** p < 0.05, *** p < 0.01; all variables are winsorised at 1-99 percent level. Industry and region effects have been included in the analysis.16

VARIABLES Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12 Model 13 Model 14

Controls Controls & independent

Controls, independent & interaction effect medium-sized board

Controls, independent & interaction effect share of independent directors

(27)

Table 7. TR on Stakeholders’ Legitimacy

Table 8. LR on Shareholders’ Legitimacy

VARIABLES Model 1 Model 2 VARIABLES Model 1 Model 2

Controls Controls & independent variable

Controls Controls & independent variable DECOUPLING 0.5780*** DECOUPLING 17.1182** F_SIZE 0.4692 *** 0.4582*** F_SIZE 5.5188*** 5.1206*** RD_INT -0.0548* -0.0677** RD_INT 1.945** 1.6049 LEV -0.1825 -0.2357 LEV -1.2735 -2.5107 SLACK 0.0475 0.0387 SLACK 0.9461 0.7425 GROWTH -0.1395 -0.1289 GROWTH -2.0113 0.7425 Constant -7.6841*** -7.4648*** Constant -73.3800*** -69.1072*** Pseudo R2 LR chi2 F-statistic 0.3089 217.94*** -243.84511 0.3206 226.20*** -239.71714 R2 Adjusted R2 F-statistic 0.2640 0.2290 7.73*** 0.2805 0.2471 8.40*** Number of observations: 249; * p < 0.10, ** p < 0.05, *** p < 0.01; all variables are winsorised at 1-99 percent level

Industry and region effects have been included in the analysis.

Table 7 displays the results for hypothesis 5. Decoupling positively and significantly impacts stakeholders’ legitimacy (b=0.5780, p<0.01), which means that H5 is supported. On top of that, firm size has a significant positive influence on legitimacy in (p<0.01). Further, R&D intensity negatively and significantly affects stakeholders’ legitimacy (p<0.05, p<0.1). The other controls have no effect.

Table 8 shows the results for hypothesis 6. Decoupling has a positive and significant impact on shareholders’ legitimacy (b=17.1182, p<0.05), implying that support is found for H6. So, decoupling increases the market value per share with $17.12. Moreover, firm size has a positive and significant relationship with shareholders’ legitimacy (p<0.01). Last, R&D has a positive and significant relationship with shareholders’ legitimacy in model 1 only (p<0.05). Other control variables have no effect.

Table 9. Overview of the Results

Sign Support p-value Comment

H1 + Yes <0.01

H2 + Yes <0.05

H3 + Partial <0.01 Pressure for environmental performance positively and significantly affects decoupling (p<0.01). However, no effect was found for earnings pressure.

H4 - Partial <0.1 <0.05

A medium-sized board and CEO/chair separation negatively moderate the effect of earnings pressure on decoupling. No other moderation effects were found.

H5 + Yes <0.01

(28)

5. DISCUSSION

The purpose of this paper was to study the effect of the two conflicting pressures for environmental performance and short-term earnings on the decoupling of innovation activities, and subsequently the effect of decoupling on legitimacy in the eyes of the firm’s stakeholders and shareholders. In addition, I wanted to look if board composition could limit executives’ opportunities to decouple.

The findings first suggest that the pressure for environmental performance increases the number of empty green patents and earnings pressure enhances R&D cuts. However, when looking at the decoupling of innovation activities, only the pressure for environmental performance increases the probability of decoupling. This may imply that the pressure for environmental performance may be more influential on executives’ decision to decouple. So, attempting to comply to environmental performance pressure may be considered as more important than earnings pressure, suggesting that firms can no longer ignore pressure for environmental performance (Hart, 2013). Further, the results show that CEO/chair separation and a medium-sized board may be necessary to limit decoupling, whereas incorporating more independent directors has no effect. Previous scholars have noted that CEOs may have considerable impact on the choice of independent directors. As a result, independent directors may represent members that are unlikely to challenge the CEO and/or are more likely to accommodate the CEO’s wishes. In addition, there may be information asymmetry between the CEO and independent directors with regard decisions about R&D and green innovation (Bebchuk, Fried, & Walker, 2002). Hence, independent directors may lack knowledge about R&D or may not have enough understanding of environmental matters to detect R&D cuts and/or the less valuable green innovations (empty green patents). So, this may decrease their ability to reduce decoupling. Further, legitimacy in the eyes of the firm’s stakeholders and shareholders is following from decoupling. Thus, decoupling enables executives to create an illusion of complying to diverging demands, sufficient to attain legitimacy.

This remainder of this section discusses the theoretical and practical implications resulting from these findings. Last, limitations and suggestions for future research are presented.

5.1 Theoretical Implications

(29)

environmental harm. Additionally, this study suggests that green innovations served for other motives than economic opportunities may have different characteristics. These innovations may merely serve as symbolic signals and may therefore lack substantive value. Although some explanations of symbolic greening had been put forth (Ramus & Montiel, 2005), it deserved more attention from literature to determine its potential causes. Moreover, the concept of empty patents had not been explored in prior studies, which was a research void this study aimed to fill.

Next, even though scholars found empirical support for REM via R&D cuts (e.g., Osma & Young, 2009), no attention had been paid to what executives do to obscure these cuts. As such, this study is the first to suggest that firms may use symbolic, externally-oriented innovation activities to mask substantive, internally-oriented cuts in R&D. So, empty patents can be the symbolic signals for innovation and long-term performance, by which they can disguise R&D cuts.

Moreover, decoupling required more empirical research to find out when it occurs and how widespread it is (Scott, 2001; Tilscik, 2010). In this regard, this study contributed to research on decoupling because it examined decoupling of an important aspect of the firm’s operations, which had not been researched before. Also, the decoupling as an attempt to simultaneously comply to mounting pressure for environmental performance and earnings pressure was a concept that had not been studied, whereas it could be common-practice in firms from all over the world.

5.2 Practical Implications

The findings of this study unfold several practical implications. First, this study provides evidence that executives engage in opportunistic behaviour–R&D cuts, empty green patenting, and decoupling–which is the result of the short-term focus of executives caused by short-term compensations plans and the pressure to achieve short-term performance (Christensen et al., 2010). To solve this, it may be important for firms to implement strategic controls as opposed to financial controls because strategic controls result in more commitment to long-term, strategic investments (e.g., in R&D) (Hitt, Hoskisson, Johnson, & Moesel, 1996). Further, it is important to restructure managerial compensation plans towards the long-term, relying on multi-year performance targets. Firms could implement equity-based compensation plans such as stock options or restricted stock grants to steer executives’ focus towards long-term performance (Core, Guay, & Larcker, 2003; Mehran, 1995). In addition, policy makers could intervene by imposing regulations that force executives to hold stock options or unsecured creditor interests, which causes them to think twice about risking the firm’s long-term performance (Walker, 2010).

(30)

environmental performance. This means that they should rethink how policies should be structured in order to stimulate firms to undertake substantive actions to increase their environmental performance. For instance, an additional standard (like an ISO code) could be developed to indicate which firms invest in the more valuable green innovations (cited green patents).

Further, policy makers and shareholders’ representatives could aim for a measure similar to the revised Dutch Corporate Governance Code of 2016, enforced to stimulate long-term value creation. Together with a change in compensation plans, such a measure may guide executives to increasingly consider long-term firm value. Next, because CEO/chair separation and a medium-sized board decrease decoupling–and have been shown to reduce opportunistic behaviour in previous studies–it would be better if they become the rule of thumb.

5.3 Limitations and Future Research

Initially, this study wished to include the entire energy sector. However, due to insufficient data availability and because some industries had to be removed to be able to run the regressions to determine levels of abnormal R&D, many firms were dropped. This may impact the generalisability of the results to the entire energy sector. Furthermore, industries with two-digit SIC codes from 44-50 have been included in the sample. These codes are generally excluded in studies of earnings management because these are highly-regulated industries. Nevertheless, as my sample was focused on the energy sector only and these industries were greatly represented in my sample, I chose to include them. However, this could have led to deviant results. For future scholars it would be interesting to do more research on the energy sector when data has gotten more sophisticated. Additionally, future research could focus on other industries that are increasingly subject to environmental performance pressure, like the automotive sector or energy-intensive manufacturing industries.

Moreover, independent directors in this study include non-affiliated as well as affiliated directors which could have affected the results. Future research could look at the effect of different types of independent directors since these could have a differential influence on limiting opportunistic behaviour. Further, the effect of other corporate governance mechanisms on decoupling deserves more attention from future research if data sufficiency has increased. For instance, an environmental committee on the board could be related to less empty patents as this committee potentially has more knowledge about the value of green innovations. However, this measure could not be included due to large data unavailability.

(31)

6. CONCLUSION

This paper aimed to empirically examine the effect of the two contradictory pressures for environmental performance and short-term earnings on the decoupling of innovation activities, and ultimately, how decoupling impacts legitimacy in the eyes of the firm’s stakeholders and shareholders. The findings first suggest that firms increase the level of empty green patents as response to the pressure for environmental performance. Second, R&D cuts are increased to cope with earnings pressure. Further, a higher likelihood of decoupling results from the pressure for environmental performance but not from earnings pressure. Last, decoupling has been shown to positively affect both types of legitimacy.

(32)

7. REFERENCES

Aguilera-Caracuel, J., & Ortiz-de-Mandojana, N. (2013). Green innovation and financial performance: An institutional approach. Organization and Environment, 26(4), 365–385.

Aiken, L. S., & West, S. G. (1991). Multiple regressions: Testing and interpreting interactions. Newbury Park, CA: Sage.

Alin, A. (2010). Multicollinearity. Wiley Interdisciplinary Reviews: Computational Statistics, 2(3), 370- 374. Anderson, C.R., & Reeb, M.D. (2004). Board composition: Balancing family influence in S&P 500 firms.

Administrative Science Quarterly, 49, 209–237.

Ashforth BE, Gibbs BW. 1990. The double-edge of organizational legitimation. Organization Science 1(2): 177-194.

Baber, W.R., Fairfield, P.M. & Haggard, J.A. (1991). The effect of concern about reported income on discretionary spending decisions: The case of research and development. The Accounting Review, 66, 818–829. Ballenberger, N., Lluis, A, von Mutius, E., Illi, S., & Schaub, B. (2012). Novel statistic approaches for non-normal censored immonulogical data: an analysis of cytokine and gene expression data. PloS one, 7(1), 1-7 (e46423).

Bange, M.M. & De Bondt, W.F.M. (1998). R&D budgets and corporate earnings targets. Journal of Corporate

Finance, 4, 153– 184.

Bansal, P. (2002). The corporate challenges of sustainable development. Academy of Management Executive, 16, 122-131.

Bansal, P. (2005). Evolving sustainably: A longitudinal study of corporate sustainable development. Strategic

Management Journal, 26, 197-218.

Bansal, P., & Gao, J. (2006). Building the future by looking to the past: Examining research published on organizations and environment. Organization & Environment, 19, 458-478.

Bansal, P., & Roth, K. (2000). Why companies go green: A model of ecological responsiveness. Academy of

Management Journal, 43(4): 717–736.

Bebchuk, L.A., Fried, J.M., & Walker, D.I. (2002). Managerial power and rent extraction in the design of executive. The University of Chicago Law Review, 69, 751-846.

Beneish, M. D. (1999). The detection of earnings manipulation. Financial Analysts Journal, 55(5), 24–36. Bereskin, F.L., Hsu, P., & Rotenberg, W. (2018). The real effects of real earnings management: Evidence from innovation. Contemporary Accounting Research, 10 (10): 1-33.

Berrone, P., Fosfuri, A., Gelabert, L., & Gomez-Meija, L.R. (2013). Necessity as the mother of “Green” inventions: Instititional pressures and environmentlal innovations. Strategic Management Journal, 34, 891– 909.

‘‘Decoupling of R&D and Empty Green Patents as Response to Conflicting demands; and Increased Legitimacy Following This’’