The effect of External Selection events on Firm Capabilities:

(1)

Master of Science in International Business and Management Master Thesis

The effect of External Selection events on Firm Capabilities:

How the Paris Agreement impacted the strategic behavior of non- renewable energy companies

Author: Deirdre Coveney Student number: S3451720 Email: d.coveney@student.rug.nl

Supervisor: Dr. K. McCarthy Co-Assessor: Dr. R. De Vries Date of Submission: 17^th of June, 2019

Word count: 14,924 (excluding abstract, tables and references)

Faculty of Economics and Business University of Groningen

Duisenberg Building, Nettelbosje 2, 9747 AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands

http://www.rug.nl/feb

(2)

(3)

Abstract

This study aimed at investigating how and why firms respond following an external selection event that has impacted their capabilities. Using the capability lifecycle determined by Helfat

& Peteraf (2003), there is existing literature that agrees that firms must adapt following external environmental changes. The mechanisms that force these firms to adapt and how they aim at regaining their capabilities were of interest for this study.

The selection event analysed was the Paris Agreement in 2015, with the sample selected from non-renewable energy companies, based on literature substantiating how susceptible these industries are to external selection events. This study aimed at exploring the effect of this regulation on firm value, in addition to considering how firms reacted following its codification, particularly in terms of how they aimed at adapting and nurturing new capabilities.

The behavioral reactions were divided into internal and external capability development.

The main results of this study revealed that firm value was initially significantly negatively impacted by the Paris Agreement, but that the selection event provided opportunities to create firm value additionally. Following the event, firms immediately intended to develop their internal capabilities, shown through increased internal spending, in anticipation of the regulation. However, in subsequent years, firms explored external capability development through Mergers and Acquisitions, as an alternative to internal development, which aligns with previous studies.

Overall, this study gives credence to the study of capability development, capability lifecycles and how firms adapt following a selection event, all of which are recommended for further research.

Key words: capabilities, capabilities lifecycle, selection event, firm value, capability development.

(4)

(5)

List of Tables, Figures & Formulae

Tables: Page:

Table 3.1: Overview of Industries within Sample 14

Table 4.1: Descriptive Statistics of all variables from the period of 24

Table 4.2: OLS Regression Results for Firm Value 28

Table 4.3: Lagged OLS Regression Results for Firm Value 29 Table 4.4: NBR Results for Merger and Acquisition Behaviour 33 Table 4.5: Lagged NBR Results for Merger and Acquisition Behaviour 34

Table 4.6: OLS Regression Results for Free Cash Flow 36

Table 4.7: Lagged OLS Regression Results for Free Cash Flow 37

Table 4.8: VIF Results of the independent variables 39

Table 4.9: Correlation Matrix of Variables 40

Table 5.1: Hypothesis Results 41

Figures: Page:

Figure 2.1: Stages of the capability lifecycle, Helfat & Peteraf (2003) 4 Figure 2.2: Branches of the capability lifecycle, Helfat & Peteraf (2003) 5 Figure 2.3: Overview of the strategic options following a Selection Event 6 Figure 3.1: Distribution of Acquirer Countries within Sample 15

Figure 3.2: Full Conceptual Model with Hypotheses 20

Figure 3.3: Distribution of Free Cash Flow data before log transformation 21 Figure 3.4: Distribution of Market Value of Firm data before log transformation 21

Formulae: Page:

Formula 3.1: Free Cash Flow Formula 17

Formula 3.2: Tobin’s Q Formula 18

Formulae 3.3, 3.4 & 3.5: Regression Formulae 21

Formula 4.1: Coefficient Transformation when Dependent Variable is log- transformed

25

Formula 4.2: Coefficient Transformation when Dependent and Independent Variables are log-transformed

25

(8)

ii

List of Abbreviations

Concepts:

- RBV: Resource Based View - FCF: Free Cash Flow

- R&D: Research & Development - M&A: Mergers & Acquisitions Statistical Analysis:

- NBR: Negative Binary Regression - OLS: Ordinary Least Squares Regression

(9)

1

1. Introduction

How firms adapt to market changes is becoming increasingly studied across a multitude of academic fields (Peng, 2002; Sirmon, Hitt, & Ireland, 2007). In this regard, how firm resources and capabilities are impacted by the market and external changes in the market is critical to comprehend. There is clear evidence that firms are affected by external events, deemed ‘selection events’ (Helfat & Peteraf, 2003; Huang, Masulis & Stoll, 1996; Kilian

& Park, 2009; Lamont, 1997; Li & Tallman, 2011). Firms create value through their competitive capabilities (Ireland, Hitt, & Sirmon, 2003), however it is imperative to know how to develop and utilise these capabilities within a changed business environment following an external event.

In this study, market uncertainty was caused by the Paris Agreement of 2015, which brought forward legally binding legislation that will enforce the reduction of green-house gas emissions to maintain global warming to below 2℃ (Wei et al, 2016; Bodansky, 2016;

Dimitrov, 2016). Companies must investigate how to develop their internal core competencies & exploit knowledge in order to survive following this selection event (McCarthy, 2018; Szulanski et al., 2016; Lin & Wu, 2010). Rather than firms following the traditional trajectory for capabilities’ lifecycles (Helfat & Peteraf, 2003), this external event changes how firms operate, resulting in their needing to develop new capabilities internally (Baregheh, Rowley & Sambrook, 2009; Berchicci, 2013; Islam, Hossain & Mia, 2018) or externally (Ahuja & Katila, 2001; Barney, 1986; Henderson & Cockburn, 1996;

Wernerfelt, 1984). However, the mechanism that encourages adaption and the adaption processes they utilise following an external selection events is an area that is understudied.

In relation to motivation, there is a clear research gap in this area, in particular the capabilities perspective and how firms react to selection events. Though some external selection event papers have been published (Helfat & Peteraf, 2003; Li & Tallman, 2011;

Polzin et. al., 2015; Szolgayova, Fuss & Obersteiner, 2012), how selection events have affected firm capabilities is understudied. The endogenous strategic behaviours that firms exhibit when adapting to exogenous shock is the main theoretical contribution. Particularly, the mechanisms by which firms acquire and develop new capabilities in order to adapt following a selection event, linking the capabilities perspective with that of the selection event, is the aim of this paper. In particular, how firm behaviour aligns with the pathways dictated by Helfat & Peteraf (2003) following a selection event is of interest. Many authors

(10)

2

have given legitimacy to the study of this area, accepting that firms must adapt their behaviour following a selection event (Cavusgil & Deligonul, 2012; Tang, 2006; Shuen et.

al., 2014; Widmaier et. al., 2007). For the sample selection, existing literature has focused on the energy sector, as they have been particularly susceptible to external regulatory changes, giving credence to the focus of this paper on such industries (Jacobs, 1986; Jensen, 1986; Jensen & Smith, 1985; McCarthy, 2016; Picchi, 1985). As a result, the following research question was developed:

Research Question: How are firms affected by, and how do firms react to, selection events in order to develop or acquire new capabilities, in particular the effect of the Paris Agreement on the M&A and innovation behaviour?

In addition to the theoretical motivations detailed above, there is added interest in terms of the international aspect of this study. With firms operating in an increasing number of markets with further risks and returns available through international expansion (Errunza

& Senbet, 1981; Gubbi, Aulakh, Ray, Sarkar & Chittoor, 2010; Johanson & Vahlne, 1977), how firms can adapt to external events and develop their capabilities is important. The process of business development and how firms can capture value has been long studied (Johanson & Vahlne, 1977; Prasad, 2004). However, within an increasingly globalised market, firms are branching further into both product and geographic market development (Helfat & Peteraf, 2003), meaning that there is a need for firms to comprehend not only how they are exposing themselves to elevated risk from their expansion, but also how they can react and adapt to an increasingly dynamic environment (Li & Tallman, 2011).

In terms of the structure of this paper, the research begins with a detailed overview of the previous literature into the related fields including selection events, strategic behaviours following these selections events, and energy policy & regulations. Based on this literature review, hypotheses are developed. In Section 3, the methodology for completing this research is outlined, including the regressions selected and how each variable will be measured. In the results section, an overview of each regression, in addition to the descriptive statistics is given. In tandem with this, the overview of robustness is given.

Section 5 will give an overview of the broad contributions of this study based on the results, in addition to a comparison of the results with theory. Limitations of the research are also given in this section. In the final section, details of the implications and conclusions are given.

(11)

3

2. Literature Overview

2.1 Background

2.1.1 Research-Based View & Sustained Competitive Advantage:

The Resource-Based View (RBV) dictates that a firm’s competitive position is based on its internal resources, as opposed to its market position (Wernerfelt, 1984). A resource is anything tangible or intangible that could be a strength or weakness of a given firm, upon which strategies are built (Barney, 1986; Wernerfelt, 1984). Firms resources include assets, capabilities, organisational processes and information that is controlled by the firm in order to invent and implement strategies (Barney, 1991).

Barney (1986) defined a sustained competitive advantage as a value creating strategy that is not simultaneously being implemented by other firms and where there is no potential for the benefits of this strategy to be duplicated, rather than a competitive advantage that lasts a long time (Jacobsen, 1988; Porter, 1985). To possess this sustained competitive advantage, firm resources and capabilities must fulfil the VRIN framework: being valuable, rare, inimitable and non-substitutable (Barney, 1991). For firms to maintain their competitive advantage, they must continuously upgrade these capabilities (Porter, 1980). Firms create value through their competitive resources (Barney, 2002; Ireland, Hitt, & Sirmon, 2003), however it is imperative to know how to develop capabilities in order to gain a new competitive advantage within a new context (Sirmon, Hitt, & Ireland, 2007). Uncertainty within a market can result in the firm needing to develop new competitive advantages in order to exploit new strategies (Sirmon, Hitt, & Ireland, 2007). The RBV has become widely accepted as an important and relevant theory in business strategy (Barney, 2001).

2.1.2 Capabilities Perspective:

Helfat & Peteraf (2003) extended the RBV to examine how capabilities would play a role in forming a firm’s sustained competitive advantage. Particularly, they focused on the importance of capabilities rather than resources, in that resources are static supplies, while the capabilities are the productive abilities to use these resources.

In terms of types of capabilities, Helfat & Peteraf (2003) define two types: operational and dynamic. The former refers to performing a task using a variety of routines to produce a significant output (Helfat & Peteraf, 2003), while the latter does not directly affect output but indirectly affects the output of operational capabilities (Teece et. al., 1997). Shuen et. al. (2014)

(12)

4

define 3 general types of capabilities: administration, operations and governance. The capabilities themselves are embedded into some combination of skilled personnel, facilities and equipment, and process and routines. However, these ordinary capabilities are often accessible through the market (Shuen et. al., 2014), which means that they can be bought or accessed externally as well as being developed internally. Thus, there is a clear link between ordinary and operational capabilities from both authors. These operational capabilities can be difficult and costly to outsource as they are developed through learning, economies of scope and absorptive capacity of the firm.

2.1.3 Capability Lifecycle:

The capability lifecycle illustrates a pattern and set of potential pathways that represent the evolution of capabilities within a firm (Helfat & Peteraf. 2003). This lifecycle is based on Wernerfelt’s (1984) link between products and resources: in the way that products have a lifecycle, it would follow that capabilities, or the ability of an organisation to perform a set of tasks using the resources, also have a lifecycle (Helfat & Peteraf, 2003). This lifecycle is not tied solely to one product, as the capability is fungible across products, meaning that a capability lifecycle can extend far beyond that of a product (Helfat & Peteraf, 2003).

This framework is aimed at explaining the outline of how capabilities are developed within a firm, in order to highlight the potential for business development into new products and markets (Grant, 2002; Helfat & Peteraf. 2003). The authors refer to three overall stages in this lifecycle.

Figure 2.1: Stages of the capability lifecycle. Source: Helfat & Peteraf (2003).

The lifecycle begins in the founding stage, where a capability is created. This stage requires an objective and a group or leader to begin the process of creating a new capability (Helfat &

Peteraf, 2003). By examining the endowments possessed within the firm (Levinthal & Myatt, 1994), the team creating the capability determines what social, human and cognitive capital

(13)

5

they possess in order to allocate the resources and inputs necessary to create the capability (Adner & Helfat, 2003; Helfat & Peteraf, 2003). Following this is the development stage. This is the stage where the capability is advanced and refined, and alternatives are examined. This development stage is highly dependent on the founding, as the capital and resources identified will determine the path that the capability development takes. The capability is improved and invested in continuously throughout this phase, based on the absorptive capacity of the firm, its coordination and its learning-by-doing abilities (Helfat & Peteraf, 2003). Finally, once the capability ceases to develop, it enters the maturity stage. Capability development is based on prior experience, the initial decisions in the process, and the choices made related to alternatives. Firms are limited by what they can do by the external environment, which constricts the paths that firms can take in order to develop capabilities (Helfat & Peteraf, 2003;

Peng, 2002; Peng, Wang & Jiang, 2008).

2.1.4 Selection events:

Figure 2.2: Branches of the capability lifecycle. Source: Helfat & Peteraf (2003).

Helfat & Peteraf (2003) defined a ‘selection event’, illustrated in Figure 2.2 above, as a point where firms encounter an event that has a strong enough impact to alter the environment within which the firms interacts, meaning they must change something in order to maximise with the capability lifecycle. In their article, the authors define two types of selection events: within or outside of the organisation, termed internal and external events. The authors denote examples of the external selection events including changes in demand, science and technology, availability of raw materials and government policy. Many other authors have highlighted how external events impact firms and how they operate internally within a changing external environment (Huang, Masulis & Stoll, 1996; Kilian & Park, 2009; Lamont, 1997; Li &

Tallman, 2011). Fritzer (2007) explains that there are two categories of selection events: those

(14)

6

that threaten to make a capability obsolete and those that provide new opportunities for capability growth/change.

Helfat & Peteraf (2003) further explain the mechanism by which firms must in turn adapt to the external selection events – the internal management have decisions that they can make in order to adapt to the exogenous selection event. This links with the article by Widmaier, Blyth

& Seabrooke (2007), who suggest the need to deal with an external change must be dealt with and adapted to internally within the firm. The focus of this study will be on the external selection events in the form in a regulation, to which firms must adapt, as they threaten to make a capability obsolete.

2.1.5 Firm reactions to External Selection events:

Helfat & Peteraf (2003) explain how selection events affect the evolutionary path of capabilities. In terms of adaption strategies, Helfat & Peteraf (2003) define six potential paths for capability transformation that firms can partake in following a selection event.

Figure 2.3: Overview of the strategic options following a Selection Event. Source: Author

Under extreme circumstances, such as a firm shutting down, firms can retire a capability completely which means they lose that capability. Alternatively, a capability utilisation can be reduced, meaning that the level of the capability degrades, referred to as retrenchment. It can arise that a capability can be transferred to a new geographic market, which is deemed as replication of a capability. Similarly, firms can redeploy the capability into a different product-

Selection Event

Internal Capability Development

Renewal

Recombination

External Capability Development

Redeployment

Replication

Capability Loss

Retirement

Retrenchment

(15)

7

market, usually requiring some type of alteration to the capability. If firms attempt to improve a capability, it indicates the renewal of a capability. Similar to this, firms can recombine the capability with existing capabilities in order to provide similar effects to capability renewal. It is clear that firms have a range of options they can apply, but the authors dictate that only 4 of the paths offer capability development: Replication, Redeployment, Renewal and Recombination. Contrastingly, retirement and retrenchment are related to the withdrawal or degradation of the capability over time (Helfat & Peteraf, 2003).

Many authors dictate that firms have two options to develop new capabilities: internal development or external development (Grant, 1991; Hall, 1990; Heeley et al., 2006). These studies align with the work of Helfat & Peteraf (2003) in that the firms can either invest within the firm or across firm borders. Renewal and recombination, which the authors deem highly related, are a form of internal capability development. Firms analyse their internal resources and inputs in order to invest in developing capabilities endogenously – by either combining their existing capabilities to develop new ones or by reinvesting into what they possess, they can fulfil this internal development following external events (Helfat & Peteraf, 2003; Li &

Tallman, 2011; Widmaier et al., 2007).

Contrastingly, firms can invest externally to acquire capabilities rather than develop them internally (Blonigen & Taylor, 2000; Hall 1987). Helfat & Peteraf (2003) dictate that firm development can take place across firm boundaries. One common mode of external development is through Merger and Acquisition (M&A) behaviour (Ahuja & Katila, 2001;

Barney 1986). Within the mode of M&A behaviour, firms can also diversify into new geographic or product markets (Ghoshal, 1987; Hitt, Hoskisson & Kim, 1997; Prahalad & Doz, 1987), which directly relates to replication and redeployment (Helfat & Peteraf, 2003). By acquiring the capabilities in order to branch into new markets, firms do not need to go through the learning and development process which Helfat & Peteraf (2003) describe. As a result, capabilities can be bought externally through the market or developed internally within the firm (Shuen et. al., 2014).

2.1.6 Energy Regulations & Paris Agreement:

Energy policies are primarily designed to ensure security of supply, diversity of sources and access to sustainable energy supplies in addition to accruing taxation revenues (DTI, 1998;

Helm, 2002).

(16)

8

The Paris Agreement of 2015 was drafted in pursuit of a ‘Goldilocks’ solution to climate change (Bodansky, 2016), being neither too weak for effectiveness nor too strong to be globally accepted and adopted. The primary addition to the Paris Agreement relative to prior legislation is that it is legally binding (Bodansky, 2016; Dimitrov, 2016; Wei et. al, 2016). Despite being anticipated, this legally binding aspect is the critical factor making this an external selection event to which incumbents must adapt. As the legislation is designed to create a market of roughly $13.5 trillion in the renewable energy sector, there will be a huge market increase for renewables and a contrasting huge impact on non-renewables (Kern & Rogge, 2016; Wei et al, 2016). All 196 countries have agreed to net-zero greenhouse gas emissions in order to maintain global warming below 2℃ in the second half of the century (Rogelj, et. al., 2016), with the private sector being a central actor in this change (Wei et. al, 2016). As a result, there is a clear impact on the value of resources and capabilities of incumbents due to extreme market pressure following the Paris agreement, making it an excellent example of how firms adapt following external selection events.

In terms of literature on policy changes, there has been a focus on renewable energy. Reuter, Szolgayová, Fuss & Obersteiner (2012) examine investments in renewable energy technologies based on the institutional context. Similarly, Polzin, Migendt, Täube & Von Flotow (2015) explored the effects of institutional context on investments into renewable energy companies, which has been echoed among other authors (Reuter et al., 2012; Zhou, Lizhi & McCalley, 2011). However, these studies have focussed on the institutional context rather than changes in terms of how firms adapt to external selection events. Thus, there is a clear gap that this study aims to examine.

2.2 Hypothesis Development:

2.2.1 Firm Value:

In terms of literature that examines the effects of selection events, the focus is mainly on financial and stock-related effects. Though there are distinct impacts on a variety of elements in the business environment (Li & Tallman, 2011), the focus has been on the national level in addition to the industry level. Many authors have explored how external selection events, including regulatory changes, stock shift and energy shocks, impact country-level stock markets and stock prices (Huang, Masulis & Stoll, 1996; Kilian, 2008; Kilian & Park, 2009).

On the industry-level, there has been some focus on strategy, more in relation to how firms can mitigate risks that arise from external selection events. Rather than examining how to react following the event, authors have examined how to protect the firm from these events and how

(17)

9

they can potentially learn following a selection event for the future (Tang, 2006; Cavusgil &

Deligonul, 2012). However, these studies have confirmed how external events directly impact the firm internally (Huang, Masulis & Stoll, 1996; Lamont, 1997; Lamont & Polk, 2002). In a meta-analysis, Karna, Richter & Riesenkampff (2016) highlighted that, in instances of exogenous instability, firms need to develop multiple types of capabilities in order to maintain firm performance, and in turn value. Thus, the cross-examination of selection events and strategic behaviour is an area that is under-studied and imperative in order for any firm to examine their possible avenues for navigating a newly defined business context.

In terms of the energy markets, authors explored the variety of mechanisms that affect energy prices and examined the selection events that cause these fluctuations (Huang, Masulis & Stoll, 1996; Helfat, 1997; Kilian, 2008; Kilian & Park, 2009; Lamont, 1997). A selection event, in this case the Paris Agreement, would directly affect both the resources and capabilities of energy incumbents, in particular their extraction capabilities which cannot be repurposed for renewable energy creation, as renewables do not need to be extracted in the same way. Thus, these capabilities lose value, which results in the firm needing to adapt in some form to this exogenous selection event.

Before analysing the reactions that follow a selection event, it is critical to empirically assess how the event affects the firms within an industry. This means that the effect of the selection event, measured by the change in market value of the firms in the chosen industries, acts as a mechanism to force firms to adapt. Based on previous studies, there is a clear negative effect on the market value of the firm (Kilian & Park, 2009; Lamont, 1997; Lamont & Polk, 2002).

This has been translated into the following hypothesis:

Hypothesis 1: Following the legitimisation of the Paris Agreement in 2015, the market value of firms within the non-renewable energy sectors decreased.

2.2.2 External Capability Development:

Following the Paris Agreement, incumbent energy companies will face a ‘selection event’

(Helfat & Peteraf, 2003), which means they need to develop new capabilities through internal/organic growth or external/inorganic growth (Barney, 1991; Grant, 1997; Prahalad &

Hamel, 1990; Wernerfelt, 1984). A substitution relationship between organic (internal) and inorganic (external) growth has been confirmed by multiple authors - firms partake in only one of the two modes of resource or capability development (Blonigen and Taylor, 2000; Hall, 1987; Hall, 1990; Heeley et al. 2006).

(18)

10

Acquisitions are a mode for purchasing a bundle of resources and capabilities in the external market in order to create a new competitive advantage (Ahuja & Katila, 2001; Barney, 1986;

Henderson & Cockburn, 1996; Wernerfelt, 1984). Acquisitions enable firms to acquire new resources and capabilities but also act as a mechanism to fuel internal innovation in the acquiring firm (Ahuja & Katila, 2001). By acquiring external capabilities, firms enable internal knowledge and capability development by updating their own knowledge (Lin & Wu, 2010), which is necessary to enhance innovation (Berchicci, 2013; Cloodt et al., 2006). Thus, these acquisitions are a mechanism of external capability acquisition following a selection event.

However, many studies (Dierickx & Cool ,1989; Lin & Wu, 2010; Parra-Requena et al., 2015;

Prahalad & Hamel, 1990) have demonstrated that internal growth, through recombination and renewal of capabilities, can be time & resource intensive. This suggests that external acquisitions of these capabilities can be a more suitable step (Ahuja & Katila, 2001; Dierickx

& Cool ,1989; Lin & Wu, 2010; Parra-Requena et al., 2015; Prahalad & Hamel, 1990). Due to this proven substitution relationship, it is assumed that a similar trend will take place as a result of this energy regulation. Such regulatory changes constitute external selection events, which suggest a need for quick responses (Li & Tallman, 2011). Bonaime, Gulen & Ion (2018) detail how policy and political uncertainty results in a drop in the M&A behaviour of firms.

Contrastingly, when severe changes in the policies occur, many authors have noted a drastic change in the environment for firms (Cavusgil & Deligonul, 2012; Huang, Masulis & Stoll, 1996; Tang, 2006). Overall, there is a gap in the literature related to these policy-enforced exogenous shocks in terms of how firms adapt their behaviour in order to develop or acquire new capabilities. Based on this, the following hypothesis was determined:

Hypothesis 2: Following the legitimisation of the Paris Agreement in 2015, incumbent energy firms increased their Merger & Acquisition behaviour.

Within merger and acquisition behaviour, there are two aspects that are relevant: redeployment and replication (Helfat & Peteraf, 2003). Redeployment of capabilities, or product diversification, is the expansion of a firm into new product markets (Hitt, Hoskisson & Kim, 1997). Replication, or international diversification, is defined as crossing borders to new geographic locations (Ghoshal, 1987; Prahalad & Doz, 1987). Based on previous studies, firms utilise diversification as a mechanism of spreading the business base to achieve growth and reduce risks (Chan-Olmsted & Chang, 2003; Hitt, Hoskisson & Ireland, 1994). Combining these, not only do firms have to choose between internal and external capability development,

(19)

11

but there are further choices between the types of M&A opportunities, and these further affect the firm. This means that, within the M&A behaviour, it is important to acknowledge the impact of diverse M&As.

2.2.3 Redeployment

In terms of the effect on M&A activity, the selection event would likely result in increased M&A behaviour, meaning that firms would potentially diversify to minimise risks. Chatterjee

& Wernerfelt (1991) explain the high levels of free cash fuel firms into new product markets, highlighting the product diversity aspect. Similarly, many authors have detailed the benefit for firms that partake in M&As across border and product markets (Barney, 1991; O’Brien et. al., 2013; Teece et. Al., 1997). In addition, Helfat & Peteraf deemed redeployment to increase the amount of activity, which aligns with previous authors. As a result, the following hypothesis was developed:

Hypothesis 2a: Following the legitimisation of the Paris Agreement in 2015, incumbent energy firms increased their Merger & Acquisition behaviour into product diverse markets.

2.2.4 Replication

In terms of replication in new geographic markets, the nature of regulatory change of the Paris Agreement is highly pertinent. While previous regulations were not legally binding, the Paris Agreement will enforce the regulation among all 196 member countries (Dimitrov, 2016).

Thus, if firms replicate their capabilities into other geographic markets, they would struggle, as so many countries are within this member group (Weil et. al., 2016). Some scholars have explored how operating in multiple markets can be used to hedge risks or create value (Boateng, Qian & Yang, 2008; Errunza & Senbet, 1981; Gubbi, Aulakh, Ray, Sarkar & Chittoor, 2010) and at times escape regulations (Prasad, 2004). However, in this case, it is unlikely firms would be able to benefit from this internationalisation. In addition, Helfat & Peteraf (2003) deemed that replication into new markets would not result in growth of capabilities, just the maintenance of the current ones. As a result, the following hypothesis was developed:

Hypothesis 2b: The legitimisation of the Paris Agreement in 2015 did not affect the International Merger & Acquisition behaviour of incumbent energy firms.

2.2.5 Internal Capability Development

As an alternative to acquiring new capabilities, firms have the potential to internally develop them through renewal and recombination in order to create value (Helfat & Peteraf, 2003). As detailed by previous authors, a firm’s ability to survive in the market and to continue to succeed

(20)

12

is dependent on its ability to develop new competitive advantages (Grant, 1991; Porter, 1980).

Similarly, in order to deal with an external selection event, firms must pursue one of the different options in order to maximise their current capabilities or to develop new ones (Helfat

& Peteraf, 2003). Innovation is defined as the multi-stage process by which firms translate knowledge into new capabilities in order to accelerate their competitive position in the market (Baregheh, Rowley & Sambrook, 2009). To develop this internal innovation process, firms must invest time and resources into capability development to create a competitive advantage, making it a costly process (Baregheh, Rowley & Sambrook, 2009; Berchicci, 2013; Islam, Hossain & Mia, 2018). As Helfat & Peteraf (2003) explain, recombination and renewal are forms of innovation in order to create a new capability from what the firm maintains internally.

In order to partake in this internal development, firms require internal resources, in particular financial resources (Helfat, 1997). Free Cash Flows (FCF) refers to a firm having cash in excess of that required to fund all projects that have positive net present values when discounted at the relevant cost of capital (Jensen, 1986). Jensen (1986) related this to strategy, in terms of how external factors impact internal cash flows of firms and how this encouraged managers to partake in takeover and merger behaviour. Following from Jensen (1986), Free Cash Flow was utilised to explore a variety of issues within strategy (Brau & Holmes, 2006; Brush, Bromiley

& Hendrickx, 2000; Chatterjee & Wernerfelt, 1991; Reuer & Miller, 1996; Thomsen &

Pedersen, 2000). In FCF literature, many authors have confirmed that high FCF allow firms to develop and grow into new, innovative markets, with FCF being one of the largest factors contributing to internal R&D and innovation (Chatterjee & Wernerfelt, 1991; Hall, 2002;

Helfat, 1997; Himmelberg and Petersen, 1994; Opler and Titman, 1994; Richardson, 2006).

Due to a selection event, firms may (?)have free cash flows, which means that they must decide how to utilise this excess cash (Jensen, 1986). Though many studies focus on this decision- making process on the individual level, this study will focus on the organisational level.

In addition, a substitution relationship between internal R&D spending through high FCF and external M&A behaviour has been established (Chen, Sun & Xu, 2016; Del Canto & Gonzalez, 1999; Jensen, 1986; Long & Ravenscraft, 1993; Szewczyk, Tsetsekos & Zantout, 1996).

Following the substitution effect explained above, this would mean that if firms partake in M&A behaviour, they would decrease their internal capability developments, which would be represented by an increase in their FCF. In terms of increased regulations (Li & Tallman, 2011), the selection event alters the value of capabilities meaning that firms must adapt quickly in order to survive, which means that the time-intensive internal resource development is not

(21)

13

expected to be effective in the case of incumbent energy companies. Finally, when examining this relationship, it is also critical to utilise the M&A count as a control variable. This is based on the fact that this excess FCF could be utilised for M&A behaviour, as a result of this substitution relationship. Based on the above, the following hypothesis was formulated as the contrasting side of the substitution effect:

Hypothesis 3: Following the legitimisation of the Paris Agreement in 2015, incumbent energy firms decreased their internal spending, demonstrated by an increase in their Free Cash Flow.

(22)

14

3. Methodology:

3.1 Data

Quantitative analysis will be completed using secondary data on the firm-level over a time span, accessed through Orbis & Eikon (formally Datastream) both of which are available through the University.

3.2 Sample

The sample selected will be stock-listed non-renewable energy firms. Many studies explored to relevance of using the non-renewable industries for examining the effects of regulations, with many highlighting the sector as highly affected by external selection events (Helfat, 1997;

Jensen, 1986; McCarthy, 2016), making this a highly relevant industry for study. Stock-listed firms have internal data related to spending more easily accessible than non-stock listed. The firms selected are from a range of industries including Oil, Gas & Coal extraction companies, as well as related industries. The industries shown in the table below were selected:

Table 3.1: Overview of Industries within Sample

These industries were selected as they would have their capabilities directly affected by the Paris agreement, as the Paris agreement was drafted to directly impact these industries (Wei et.

al., 2016). The below Standard Industrial Classification (SIC) codes were utilised to identify the industries within Orbis, while the corresponding North American Industry Classification System (NAICS) codes were used for Eikon. The original sample from Orbis had 958 firms SIC

Codes

NAICS Codes Description of Industry Numbers of Firms

in Sample

1221 2121 Bituminous coal & lignite surface mining 31

1222 2121 Bituminous coal underground mining 2

1241 2131, 2389 Coal mining services 5

1311 2111 Crude petroleum and natural gas 203

1321 2111 Natural gas liquids 1

1381 2131 Drilling oil and gas wells 17

1382 2131, 5413 Oil and gas fields exploration services 7

1389 2131, 2371, 2389 Oil and gas field services, not elsewhere classified 26

2911 3241 Petroleum refining 21

4612 4861 Crude petroleum pipelines 9

4613 4869 Refined petroleum pipelines 3

(23)

15

while the Eikon sample had 1371, which was then matched with Eikon’s financial data on the above industries. This left 324 firms that participated in M&A behaviour that also published the necessary data in order to measure the dependent variable. These firms are spread across 43 countries and across all the above industries, with the following distribution:

Figure 3.1: Distribution of Acquirer Countries within Sample. Source: Author

This sample contains a proportional distribution across the industries in the population, with the predominant industry involved in M&A behaviour being Crude Petroleum and Natural Gas.

Similarly, the M&As were heavily concentrated in the United States, Canada, China, United Kingdom and Australia, where 968 of the M&As in the sample occurred. The acquired firms varied much wider across sectors and countries, with the acquired firms being within 157 industries and ranging across 75 countries.

3.3 Independent variable Paris Agreement

In order to examine how the Paris Agreement, the selection event in this study, affected firm behaviour, the dependent variables will be studied over a time period. Many studies have considered the impact the regulatory changes make on the stock market, with many showing how the energy industry is particularly affected (Kilian, 2008; Kilian & Park, 2009; Lamont, 1997). The Paris Agreement was enacted in 2015, meaning in order to examine a change in

(24)

16

behaviour of firms, the use of a time-period means this difference in behaviour can be explored.

The time period selected was from 2010 to the end of 2018. This is a similar process used by other authors using panel analysis (McCarthy, 2016; Kilian & Park, 2009). This is to enable this study to examine the basic behaviour of firms before the Paris Agreement, investigate their behaviour directly before its enactment, and their behaviour in the years following, its introduction.

In order for this to be coded within STATA, an indicator variable was created, designated ‘Paris regulation’. This variable equalled 1 once the regulation was introduced, during 2015, and 0 before the regulation. As a result, the years of 2010 to 2014 were deemed ‘pre-regulation’ while 2015 to 2018 were designated ‘regulated’. This was then utilized to estimate if, following from 2015, whether there was a change in the behaviour.

3.4 Dependent Variables Firm Value

Based on the literature explained in this paper, the effect of the Paris Agreement on the value of the firm will also be analysed. Many studies have examined the effect of external selection events on firm value (Huang, Masulis & Stoll, 1996; Kilian & Park, 2009; Lamont, 1997;

McCarthy, 2016). This will be measured using the Market Capitalisation Value data from Eikon. This variable represents the sum of market value for all relevant issue level share types.

It is calculated by multiplying the requested shares type by the last closing price for the shares.

This was variable was measured on the final day of each of the year between 2010 and 2018 in order to analyse how the regulation affected its value.

Free Cash Flow

Though R&D data is a common proxy for internal growth, at the time of this study, R&D data was unavailable in the industries of interest. This is likely due to the nature of the industries of interest, as these traditional industries often do not have explicit R&D budgets though they do invest in such areas (Moncado-Paternò-Castello, Ciupagea, Smith, Tübke & Tubbs, 2010).

As a result, free cash flows will be utilised as a proxy for R&D, based on previous authors establishing this relationship (Chatterjee & Wernerfelt, 1991; Del Canto & Gonzalez, 1999;

Helfat, 1997; Richardson, 2006). This data was also accessed through Eikon. Gibbs (1993) measured FCF in relation to corporate governance and portfolio restructuring. In terms of the measures that the author chooses, Gibbs (1993) uses multiple measures of FCF, with Operating

(25)

17

Cash Flows being highly relevant. For this study, formula 3.1 was used to calculate Free Cash Flows:

Formula 3.1: Free Cash Flow Formula

𝐸𝐵𝐼𝑇𝐷𝐴 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠

Within Eikon, EBITDA is explained as Earnings before Interest and Taxes for the fiscal year plus the same period's Depreciation, Supplemental, Amortization of Acquisition Costs, Supplemental and Amortization of Intangibles, Supplemental. Total assets refer to the total assets reported on Eikon, which represents the total assets of a company. The selected industries were examined in terms of how many firms had published such data, and from this, there was a match between 325 firms whom had published such financial data and had announced M&A behaviour. The timespan of 2010 to 2018 was also utilised to examine the change in FCF over time.

Merger & Acquisition

As previously explained, the examination of M&A behaviour is highly supported (Ahuja &

Katila, 2001; Lin & Wu, 2010; Parra-requena et. al., 2015) following an external selection event (Tang, 2006; Huang, Masulis & Stoll, 1996). The data for M&A behaviour was downloaded from Orbis, with a total of 2342 observations exhibited between 2010 and 2018.

These were distributed over 958 firms, ranging from 4% to 100% acquisition shares. Firms without financial data publicly available were removed, with a total number of 1142 mergers

& acquisition made over the 9-year period. The number of acquisitions made within each year were totalled for each firm in order to examine how the behaviour changed over the timespan selected, similar to previous authors examining M&A behaviour (Blonigen & Taylor, 2000;

McCarthy, 2016).

Product Diversity

As explained in the previous section of this study, M&As into product diverse areas can be highly lucrative and constitute one of the mechanisms of capability development - redeployment (Chatterjee & Wernerfelt, 1991; Helfat & Peteraf, 2003; Hitt, Hoskisson & Kim, 1997; Lamont, 1997; O’Brien et. al., 2013). Due to the panel nature of this study, the comparison between the SIC code of the acquirer and the acquired firm for each merger is done for each year, similar to previous authors (Blonigen & Taylor, 2000; Richardson, 2006). This was then used to create an interaction term between mergers and product diversity, to measure

(26)

18

the number of M&As into new diverse product markets per company per year (Helfat &

Peteraf, 2003).

Geographic Diversity

Based on existing literature, international M&As also represent an important element in M&A studies and for capability development through replication (Dos Santos, Errunza & Miller, 2008; Errunza & Senbet, 1981; Gubbi et. al., 2010; O’Brien et. al., 2013). In order to analyse geographic diversity, a similar procedure is used as above. A comparison between the home country of the acquirer and the acquisition country is made, with 1 being all M&A behaviour being completely geographically diverse and 0 being no geographic diversity in all the M&A behaviour in that specific year. This was then used to create an interaction term between mergers and international diversity, to measure number of M&As into new diverse geographic markets per company per year (Helfat & Peteraf, 2003).

3.5 Control variables Tobin’s Q

Following from the use of Gibbs (1993) measurement of Free Cash Flow, Tobin’s Q can be another measure of FCF, and thus making it a valid control variable. Tobin’s Q also represents the whether the firm makes profitable investments or make poor investment decisions (Brush, Bromiley & Hendrickx, 2000; Dos Santos et. al., 2008; Richardson, 2006), which might influence their behaviour. Tobin’s Q was calculated using formula 3.2 (Gibbs, 1993):

Formula 3.2: Tobin’s Q

𝑀𝑎𝑟𝑘𝑒𝑡 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑆𝑡𝑜𝑐𝑘 + 𝐷𝑒𝑏𝑡 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠

For this, the market value of stock data was measured by the Company Market Capitalisation variable on Eikon, which is defined as represents the sum of market value for all relevant issue level share types. The debt was measured using Total Debt outstanding. The Total Assets were measured using the same variable as used for FCF. The Tobin’s Q value for each firm was calculated for each year based on the above formula using Excel. Tobin’s Q was not uniformly distributed, meaning it was then logarithmically transformed in order to utilise regression analysis.

(27)

19 Firm Size

The second control variable included is the firm size, as larger firms will have more time and resources to invest into M&A and innovation (Blonigen & Taylor, 2000; Dos Santos et. al., 2008; Puranam & Srikanth, 2007). Many studies use firm size as a control variable, in particular when analysing M&A and financial behaviour (Ahuja & Katila, 2001; Lin &Wu, 2010). Firm size was measured by the logarithmic transformation of the number of employees over the time span, and this data was gathered from Eikon for the corresponding firms in the sample. This measure is frequently used in M&A literature (Ahuja & Katila, 2001; Lin &Wu, 2010; Puranam

& Srikanth, 2007).

Country

In addition, controls will be established for country-level effects. This is possible using the countries within which the firms are located, which can be controlled for on STATA. Country- level effects are often controlled when analysing firm behaviour and selection events (Adner

& Helfat, 2003; Cavusgil & Deligonul, 2012; Lamont & Polk, 2002).

Debt

In addition, debt is used as a control variable. Many authors studying cash flows and M&A behaviour following selection events have controlled for debt (Blonigen & Taylor, 2000;

Richardson, 2006). Firms with more debt may influence their use of cash flows, so debt is analysed using the Debt data accessed through Eikon. Similarly, as Gibbs (1993) analysed debt in terms of financial leverage, it is important to control for this. Debt is also logarithmically transformed, to ensure it is normally distributed.

Industry of Acquirer

Wilcox, Chang & Grover (2001) highlighted that acquirer industry is important to control for when analysing M&A behaviour. Similarly, other authors examining diversification effects also controlled for the acquirer’s SIC code or by only using one industry to limit the effect of various industries (Blonigen & Taylor, 2000; Hitt, Hoskinsson & Kim, 1997; King, Slotegraaf

& King, 2008). Thus, it was important to control for the individual industries within the sample.

As a result, the acquirer’s SIC codes were also controlled for in each regression model.

(28)

20 3.6 Full Conceptual Model

Based on the integration of the independent variables with the dependent, and the inclusion of control variables, the following conceptual model was created to illustrate the relationships.

Figure 3.2: Full Conceptual Model with Hypotheses. Source: Author

3.7 Statistical Technique

In order to analyse how this external event affects the behaviour of firms, a regression is necessary to understand the relationship between the independent and the dependent variables.

Firstly, the data gained from the databases was collated into one spreadsheet, matched using Excel, and organised in a Panel. This was then analysed using STATA in order to gain insights into the relationships between the above variables.

In terms of the regressions used, the nature of the data is critical to understand. The dependent variables of Firm Value and Free Cash Flow are both continuous variables, meaning that Ordinary Least Squares regression (OLS) is a suitable approach. However, one of the critical assumptions of OLS regression is that the data is normally distributed. This means that if the conditional variance exceeds the conditional mean, then the regression will not give significant results. There are a variety of ways this can be demonstrated, including creating a histogram of each variable. This method of checking for normality is selected, as the panel aspect of the data makes detecting departures from the normal very difficult with traditional means (Alejo,

H2b - H2a + H2 +

H1 -

H3 +

Independent Variable Paris Agreement

Regulation

Market Value of

Firms Product

Diversity

Tobin’s Q Firm Size

Debt Country

Year Acquirer

Industry

Control Variables Dependent Variables

Merger &

Acquisitions Free Cash

Flow

Geographic Diversity

(29)

21

Galvao, Montes-Rojas & Sosa-Escudero, 2015). This was done for each variable in question, and the results can be seen below:

Figure 3.3 & 3.4: Distribution of Free Cash Flow data and Market Value of Firm data before log transformation. Source: Author

As a result of the non-normal nature of the data, both the data for FCF and Market Cap were logarithmically transformed. This resulted in the data being normally distributed, meaning that it could be analysed using OLS.

For the M&A data, a different type of regression was required. This is because M&A data is count in nature. This means that OLS, Negative Binary regression (NBR) and Poisson regression are possible for the data. However, due to the panel aspect of this study, there are many zeros are present in the panel, meaning that the data is over-dispersed. This means that the data is also not normally distributed. As a result, NBR is more suitable as it accounts for non-normally distributed data.

Based on the above, the full regression formulae are as follows:

Formulae 3.3, 3.4 & 3.5: Regression Formulae

FIRMVALUE𝑖𝑡 = 𝛽1REGULATION𝑖𝑡 + 𝛽2M&A𝑖𝑡 + 𝛽3TOBINSQ𝑖𝑡 + 𝛽4DEBT𝑖𝑡 + 𝛽5SIZE𝑖𝑡 + 𝛽6INDUSTRY𝑖𝑡 + 𝛽7YEAR𝑖𝑡 + 𝛽8COUNTRY𝑖𝑡 + 𝜀𝑖𝑡

FREECASH𝑖𝑡 = 𝛽1REGULATION𝑖𝑡 + 𝛽2FIRMVALUE𝑖𝑡 + 𝛽3M&A𝑖𝑡 + 𝛽4TOBINSQ𝑖𝑡 + 𝛽5DEBT𝑖𝑡 + 𝛽6SIZE𝑖𝑡 + 𝛽7INDUSTRY𝑖𝑡 + 𝛽8YEAR𝑖𝑡 +

𝛽9COUNTRY𝑖𝑡 + 𝜀𝑖𝑡

M&A𝑖𝑡 = 𝛽1REGULATION𝑖𝑡 + 𝛽2TOBINSQ𝑖𝑡 + 𝛽3DEBT𝑖𝑡 + 𝛽4SIZE𝑖𝑡 + 𝛽5INDUSTRY𝑖𝑡 + 𝛽6YEAR𝑖𝑡 + 𝛽7COUNTRY𝑖𝑡 + 𝜀𝑖𝑡

(30)

22

Where FIRMVALUE is the Firm Value of firm i in the year t, FREECASH stands for the Free Cash within the firm i in the year t, and M&A represents the M&A count of firm i in the year t. The term REGULATION is the created indicator variable that denotes the Paris Agreement.

The control variables are firm’s Tobin’s Q, debt, size, industry, country and the year. The last term, 𝜀, denotes the error term.

In addition, lagged regression will be analysed. Within panel data, time variants such as lagged analysis are very useful in analysing events and their implications. Time lags account for the fact that it can take time for certain events to have an exhibiting effect on the dependent variable. While the primary regression looks at the effect of the regulation on each dependent variable in one time period, the lagged analysis analyses the effect of the regulation in a later time period, which allows for a delay in exhibited reaction to the regulation. As a result, regressions were completed using the same independent variables with time lags. The choice of the time lag is highly debated (Dormann & Griffin, 2015), but for this paper 1-, 2- and 3- year time lags were selected. As the Paris Agreement was legitimised in 2015, there have been three years following this, meaning that a choice of 3 years was logical with the total dataset running until the end of 2018.

3.8 Validity:

Validity refers to how valid or true a measurement represents accurately the features of a phenomenon that it is trying to explain or describe (Winter, 2000). Within quantitative research, this refers to how accurate the measurement aspect is. Within validity, there is internal and external validity.

3.8.1 Internal validity:

This refers to whether the results of the research relate to or are caused by the phenomenon being studied, rather than other causes (Winter, 2000). As a result, the use of certain proxies for variables should be examined. Most of the proxies selected are standard in Strategic research, however one element for study is the use of Free Cash Flow data rather than R&D data. Though R&D data can itself be useful, for industries such as the sample analysed here, the lack of accessible data is likely based on the nature of the industries studied (Moncado- Paternò-Castello, Ciupagea, Smith, Tübke & Tubbs, 2010). Thus, FCF is a proxy utilised by many authors (Chatterjee & Wernerfelt, 1991; Del Canto & Gonzalez, 1999; Helfat, 1997;

Richardson, 2006), in particular when analysing R&D.

(31)

23 3.8.2 External Validity:

The external validity of the study relates to how generalisable the results are outside of the sample itself (Winter, 2000). Thus, though this study’s results will initially represent the effects of energy regulation on energy incumbents, the industries selected were chosen based on the substantial evidence that energy industries are highly affected by external selection events (Helfat, 1997; Huang, Masulis & Stoll, 1996; Jensen, 1986; Kilian & Park, 2009; Lamont, 1997; Li & Tallman, 2011; McCarthy, 2016). Thus, the use of these industries permits the examination of how firms adapt to create capabilities following section events, rather than simply how energy firms can transition following a selection event. Consequently, the results should enable cross-utilisation across fields.

(32)

24

4. Results

4.1 Descriptive Statistics:

As previously detailed, the original sample included a total of 956 firms across 68 countries.

The financial data resulted in 1371 firms across 75 countries. Once combined, the sample for this study totalled 324 firms across 43 countries. In terms of reliability, the countries within the sample were well represented. Industries within the sample were also well distributed.

An overview of the descriptive statistics can be seen in Table 4.1 below. As previously described, various variables were logarithmically transformed in order to enable OLS regression to be utilised. The redistribution of the values ensured that the maximum was reduced, which aided in the regression analysis in terms of outliers. These variables included Firm Value, Free Cash Flow, Tobin’s Q, firm Debt and number of employees, to ensure that the variables are normally distributed.

Table 4.1: Descriptive Statistics of all variables from the period of 2010 to 2018

(1) (2) (3) (4) (5)

VARIABLES N Mean Standard

Deviation

Minimum Maximum

Regulation 2,925 0.444 0.497 0 1

Firm Value 2,925 9.683e+07 3.397e+08 0 4.387e+09

- Log Firm Value 2,693 16.10 2.513 -1.949 22.20

Free Cash Flow 2,925 -0.00380 4.308 -229.4 0.930

- Log Free Cash Flow 2,460 -2.269 0.797 -8.322 -0.0729

Merger & Acquisition Count 2,925 0.390 0.760 0 9

Internationally Diverse Merger & Acquisitions 2,925 0.226 0.615 0 9

Product Diverse Merger & Acquisitions 2,925 0.128 0.442 0 6

Tobin’s Q 2,925 6.429 157.0 0 5,511

- Log Tobin’s Q 2,685 -0.00342 0.731 -4.649 8.615

Firm Debt 2,925 32,789 100,195 0 1.321e+06

- Log Firm Debt 2,558 8.458 2.486 -4.164 14.09

Firm Size 2,925 9,185 44,480 0 552,810

- Log Firm Size 2,394 6.669 2.430 0 13.22

Year 2,925 2,014 2.582 2,010 2,018

Industry 2,925 1,538 720.0 1,221 4,613

Company 2,925 163 93.84 1 325

Regulation L1 2,600 0.375 0.484 0 1

Regulation L2 2,275 0.286 0.452 0 1

Regulation L3 1,950 0.167 0.373 0 1

4.2 Interpretation of Regression Results

There are two aspects that must be understood before the regression results can be examined in detail.

(33)

25

Firstly, the use of NBR is highly useful for count data. However, when interpreting the results, it is imperative to transform the coefficients in order to interpret the results correctly. The coefficients given directly from the NBR represent the change in the log-count of the dependent variable for a unit change in the independent (Hilbe, 2011). As a result, the regression coefficients are exponentially transformed, to give the Incidence Rate Ratios to give equivalent outputs to the OLS regression. This conversion also means that interpretation differs, in that any value below 1 is negative, while any value above 1 is positive. Thus, the Tables 4.4 and 4.5 demonstrates the result of the regressions with the coefficients transformed.

Secondly, as the some of the independent and dependent variables have been logarithmically transformed, there is two conversions that must be made in order to interpret the regression coefficients. To convert the coefficient when the dependent variable log-transformed, formula 4.1 was utilised (Benoit, 2011; Yang, 2012), where 𝛽 is the coefficient computed from the regression analysis:

Formula 4.1: Coefficient transformation when Dependent variable is log-transformed

𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝐷𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑖𝑠 𝑙𝑜𝑔 − 𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑑

= 𝑒^𝛽− 1

When both the independent variable and the dependent variable are log-transformed, then a 1%

change in the independent is explained by the following change in the dependent variable (Benoit, 2011; Yang, 2012), as seen in formula 4.2:

Formula 4.2: Coefficient Transformation when Dependent & Independent variable are log-transformed

𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑏𝑜𝑡ℎ 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 𝑙𝑜𝑔 − 𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑑

= 100(1.01^𝛽− 1)

To ensure that all the tables are cohesive, the relevant conversions have been made in the subsequent tables.

4.3 Regression Results

As there are three overall dependent variables within this study, each will first be analysed individually. Following from the primary regressions, lagged regressions with a 1-year, 2-year and 3-year lag are done and are analysed for each of the dependent variables.

(34)

26 Firm Value:

The results from the OLS regression analysis can be seen in Table 4.2. Two of the models, models 2 and 3, analysis the impact of the Paris Agreement regulation on the firm value. Model 1 is included in order to analyse without this regulation included, for comparative purposes.

Model 1 represents the relationship without the regulation or controlling for the firm’s M&A count. This means that the difference between the inclusion of the regulation and M&A count can be seen separately. As seen by the coefficients in Model 1, all of the control variables were highly significant, except for Industry. In fact, Tobin’s Q, Debt and Firm size all had positive, though small, influences on the firm value, with only the year having a negative influence.

However, when comparing with model 2, the regulation is highly significant in having a negative influence on firm value. As seen in Table 4.2, the inclusion of the regulation drastically changes the results of the regression. In addition, all control variables in Model 2 are seen to be significant except for Industry and Year. In both models, Tobin’s Q, debt and firm size have a positive effect on firm value.

For Model 3, the regulation and M&A count are both integrated. In this model, the regulation has a slightly less negative influence on firm value, at 39.8% rather than 39.5% in Model 2, but it is clear that the regulation persistently negatively influences firm value. The M&A count continues to have a positive, significant influence on firm value, and as the M&A count increases by 1, the firm value increases by 17.9%. Similar to previous models, Tobin’s Q, debt and firm size are all significant. The year and industry remain insignificant, similar to Model 2, though both with a positive, if small, effect.

Lagged Analysis on Firm Value:

In Table 4.3, when analysing the regulation with a 1-, 2- and 3-year lag, there is a clear distinction between the regression results in time t, relative to time t+1, t+2 and t+3. From Table 4.2, it was clear that the regulation had an overall negative affect on the firm value.

However, in Table 4.3, the lagged analysis demonstrates that the regulation had a positive impact. In Model 1, the regulation is only slightly significant. The M&A count is highly significant with a positive effect on firm value. The overall trend in the lagged results will be further analysed in the Discussion section of this paper

In Model 2, the firm value is even more positively affected by the regulation which is highly significant. In Table 4.2, the regressions were run to examine the effect of the regulation on

(35)

27

firm value in that year. However, Table 4.3 looks at these regulatory effects with a delay, meaning that there is a clear distinction between there being an immediately negative effect, which changes when these effects are examined with a lag. Similar to the previous models, all of the control variables, except for industry, remain highly significant. In Model 2, the addition of the M&A count has a similar effect as with the 1-year lag, in that the overall regulation has an even further positive effect relative to the first Model.

In Model 3, there is a decrease in the positive effect of the regulation on firm value though there is no statistical significance. Once again, all of the control variables of both Models remain highly significant, except for industry. Consistent with the previous models, the M&A count remains positive and highly significant, which will be discussed in the Discussion section of this paper. Across all three models, the M&A count increases with the time-lag and remains highly significant.

The effect of External Selection events on Firm Capabilities:

Master of Science in International Business and Management Master Thesis