Master Thesis
The Influence of Board Diversity and Firm Internationality on
Corporate Cost Behavior
By
Jannik J. Foertsch S3057240
Supervisor: Dr. Peter P.M. Smid
MSc International Financial Management Faculty of Economics and Business
governance and corporate cost behavior. Using a sample of 528 industrial firms, listed at the New York Stock Exchange, this research hypothesizes that an increase in board diversity along the dimensions age and gender improves corporate governance functions and abilities, consequently reducing the firms’ asymmetric cost behavior. Further, it is examined whether firms with a higher degree of internationality show less asymmetric cost behavior.
Contradicting the expectations, the results of this study indicate cost anti-stickiness for cost of goods sold. Further, results show gender diversity to improve cost behavior, while firm internationality increases cost stickiness. These findings help evaluate corporate governance practices and improve understanding the optimal characteristics and composition of a corporation’s board.
2. LITERATURE REVIEW 4
2.1.COST STICKINESS 4
2.2.COST STICKINESS AND MANAGERIAL INFLUENCES 7 2.3.OPPORTUNISTIC MANAGERIAL BEHAVIOR 9
2.3.1.EMPIRE BUILDING 10
2.4.CORPORATE GOVERNANCE 11
2.3.1.THE ANGLO-SAXON CORPORATE GOVERNANCE MODEL 12
2.5.BOARD DIVERSITY 13 2.5.1.AGE DIVERSITY 14 2.5.2.GENDER DIVERSITY 15 2.6.DEGREE OF INTERNATIONALITY 17 3. METHODOLOGY 19 3.1.MODEL 19 3.2.DEPENDENT VARIABLE 21 3.3.INDEPENDENT VARIABLES 21 3.4.CONTROL VARIABLES 22 3.5.DATA COLLECTION 23 3.6.DESCRIPTIVE STATISTICS 24 4. RESULTS 26 5. ROBUSTNESS TESTS 32
6. CONCLUSION AND DISCUSSION 35
1. Introduction
Understanding cost behavior is one of the cornerstones of successfully managing a firm. In order to understand cost behavior, researchers have proven the traditional cost and
management accounting theory, which implies a linear relationship between business activity and costs, wrong. Moreover, a considerable amount of studies depict cost behavior to be asymmetric: costs do not change proportionally to business activity (Anderson et al., 2003; Anderson and Lanen, 2009; Chen et al., 2012; Subramaniam and Weidenmier, 2003).
Accordingly, a 1 percent increase in business activity does not result in a 1 percent increase in variable costs, and vice versa. Further research examined agency problems and their
implications towards corporate governance (Carter et al., 2007, 2003; Mishra and Jhunjhunwala, 2013a). The scholars not only postulate a mitigating effect of corporate governance on asymmetric cost behavior, but moreover posture board diversity to improve corporate governance functions.
This thesis explores the relationship between corporate governance, the degree of firm internationality and increased board diversity along the dimensions age and gender, and cost behavior.
The preceding is of interest and importance, as the adjustment of the operational leverage to economic and environmental circumstances constitutes a vital element of management’s area of responsibility and action. A factor that further increased the topic’s acuteness and
potentially complicated matters is globalization. Globalization, as a process of increasing economic interdependence of national economies across the globe through a rapid increase in international movement of goods, services, technology and capital, has led to an increasingly emerged global marketplace (Joshi, 2009; Riley, 2003). Thus, through this expanding, intertwined network of trade, economies, and political relations, national economies developed a considerable dependence on economies all over the world. As a result, consequences of international economic booms and busts have a growing impact on multinationally as well as nationally active firms.
Next to prominent examples such as business software giant SAP, who has cut about 2.250 employees or German carmaker BMW, who announced 8.000 employees to be let go in 2007, the automotive industry in the U.S. alone is estimated to have cut almost 200.000 jobs
between 2006 and 2010.
The phenomenon of asymmetric cost behavior represents a dilemma for managers who are confronted with changing economic conditions, especially when facing a decrease in operational activity. A decision is to be made regarding the ability to adjust the level of committed resources representing a decrease in the production capacity or leaving the committed resources untouched, resulting in operating with unutilized production capacity. This decision depends on multiple factors, such as the manager’s expectations concerning future operational activity, the permanence of the lower activity level, personal characteristics such as risk preference, or the strength and features of corporate governance mechanisms (Balakrishnan and Gruca, 2008; Chen et al., 2012; Papadakis et al., 1998).
The goal of this study is to determine if corporate governance characteristics in form of board diversity as well as a higher degree of internationality can contribute to reducing cost
stickiness. Hence, the results will be of value to researchers and practitioners, as the results are expected to provide information regarding corporate governance characteristics that impact cost behavior.
Weiss (2010) showed in his empirical examination that firms with more asymmetric cost behavior have less accurate analysts’ earnings forecasts. Accurately predicting cost behavior hence is an essential element of earnings predictions and firm-valuation (Weiss, 2010). Contributing to the further understanding of this matter thus will valuably complement the ability to precisely predict earnings with manifold beneficiaries. Further, these contributions will provide important implications for accountants and other professionals who deal with and evaluate cost changes in association to changes in revenues, and moreover supply research on corporate governance with valuable information.
As an act of improving the governance of firms, scholars research possible improvements of corporate governance, board diversity being one of them. Board diversity receives growing attention, especially since multinational companies are increasingly expected to be governed by a board that at least partially reflects the firm’s various demands, challenges and
Beginning, this research explores if asymmetric cost behavior prevails for the sample of US-American firms.
Hence, the first research question is:
To what extent, if at all, is the cost behavior for US-American industrial firms asymmetric?
Further, this study aims to determine if an increase in the board of directors’ diversity along the dimensions age and gender leads to a decrease in the asymmetry of the sample firms’ cost behavior through an improvement in corporate governance abilities.
Accordingly, the second research question is:
Does an increase in board diversity along the dimensions age and gender lead to less asymmetric cost behavior?
Finally, in order to examine the effect of a firm’s degree of internationality on the aforementioned relationship, the third research question is:
Does a higher degree of internationality lead to a lower degree of asymmetric cost behavior?
In order to answer these questions, a regression analysis on the dataset of 528 industrial firms, listed at the New York Stock Exchange, will be run, as further explained in the section
addressing the research methods.
This study continues with relevant literature regarding cost stickiness and opportunistic managerial behavior. Next, literature regarding corporate governance, board diversity as well as internationality will be eamined. In the following section, the research method will be discussed, where after the results will be laid out. Following, the results of the robustness test will be exhibited. The final chapter presents the conclusions as well as limitations and
2. Literature Review
In the following chapter, relevant literature regarding cost stickiness will be presented first. Next, opportunistic managerial behavior on the basis of the agency theory and empire building will be discussed, followed by corporate governance and respective board diversity features. Finally, the literature review closes with pertinent literature on the degree of corporate internationality. Respectively, the corresponding hypotheses will be presented.
2.1. Cost Stickiness
Originating from the German literature (Brasch, 1927; Strube, 1936; Busse von Colbe, 1958), cost stickiness or cost remanence (German: Kostenremanenz) has increasingly been a topic of academic research. The traditional cost and management accounting theory implies a linear relation between the levels of activities and costs: fixed costs are considered to be
independent of the business activity while variable costs proportionally change with changes in the activity driver, independently from the direction of activity changes (Noreen, 1991). This symmetry between costs and activity presumes a 1 percent increase in activity to result in a 1 percent increase in variable costs and vice versa.
The topic of cost behavior was taken from theoretical discussion to empirical research by Anderson et al. (2003), using a large cross-sectional sample exploration of cost behavior for the first time. By doing so, the scholars not only implemented a methodological precedent for many following studies, but moreover introduced the “modern era” of academic research in the field of cost behavior. Anderson et al. (2003) suggest a non-linear relationship between the levels of business activities and costs, depending on whether the activity is increasing or decreasing. This type of cost behavior is labeled as “sticky”. According to the scholars “costs
Cost remanence literature examines several cost measures. Empirical research showed cost stickiness for selling, general & administrative (further: SG&A) expenses and came to find out, that SG&A costs increase on average 0.55 percent per 1 percent increase in sales whilst a decrease in sales of the same amount only leads to a decrease in SG&A costs of 0.35 percent (Anderson et al., 2003).
Another measure, the operating costs, reports an increase of costs of 0.97 percent with a respective decrease of 0.91 percent and thus confirms the findings of Anderson et al. (Calleja et al., 2006). The measure of cost of goods sold (further: COGS) shows similar results (Subramaniam and Weidenmier, 2003; Yasukata, 2011). Scholars added the measure of advertisement costs and research and development costs and found a non-linear relationship in the cost behavior as well (Banker et al., 2014). While Anderson and Lanen (2009) find empirical evidence of cost stickiness for labor costs, they did not find asymmetrical cost behavior for research and development expenses, property, plant and equipment costs, and advertising costs.
Academic research has found asymmetric cost behavior for specific industries, such as Noreen & Sonderstrom (1997) for hospitals and Balakrishnan et al. (2004) for physical therapy clinics. Submaramiam and Weidenmier (2003) depict SG&A cost stickiness for several industries, with manufacturing exhibiting the highest level of cost stickiness. In similar vein, merchandising and service firms also show asymmetric cost behavior, but not financial firms (Subramaniam and Weidenmier, 2003).
Anderson et al. (2003) as well as Subramaniam and Weidenmier (2003) describe the level of cost stickiness to be determined and influenced by the firm’s surrounding economic
conditions as well as firm and industry characteristics. These are constituted as high asset and inventory intensity, which explains the aforementioned findings of Submaramiam and
Weidenmier (2003), as well as the degree of profitability, since a competitive environment demands firms to react and adapt quickly to changing conditions.
Further, Subramaniam and Weidenmier (2003) find that the stickiness of SG&A costs and COGS depends on the magnitude of activity changes as a driving force behind the
In the literature of cost stickiness not only different industries and measures have been
examined, but also different countries. Although most studies focus on US companies, Calleja et al. (2006) conduct an empirical comparison between countries with the result that costs of German and French firms are more sticky than costs of US and UK firms.
Scholars conclude this result to be caused by differences in code-law countries and common-law countries. The main differences in the corporate governance systems that affect cost behavior are shareholder-orientation (common-law) and stakeholder-orientation (code-law). Consequently, firms are pressured by externals to take decisions in the interests of
shareholders in common-law countries, whilst firms in code-law countries are directed to take external as well as internal pressures and interests into account (Calleja et al., 2006).
Further, as cost behavior has been researched in several specific countries, studies were conducted for Italy, finding cost stickiness only for total cost of labor (Dalla Via and Perego, 2014), Australia and Argentina, amongst others. Bugeja et al. (2015) more recently proved the existence of asymmetric cost behavior for a large sample of Australian firms, although less asymmetric than for US American firms. Argentina, representing an emerging economy, showed a similar but more pronounced magnitude in cost stickiness in comparison to previous studies in emerging as well as emerged economies (Werbin and Porporato, 2012). De
Madeiros and Costa (2004) confirm asymmetric cost behavior for Brazilian firms. Scholars did not only find cost stickiness for SG&A expenses for Indonesian firms listed in the Indonesia Stock Exchange, but moreover found that the degree of cost stickiness impacts the future profitability of the firms. More precisely, an increase in asymmetric the cost behavior leads to a decrease in future profitability (Warganegara and Tamara, 2014). These findings furthermore empirically emphasize the importance of understanding cost behavior. Applying the insights of the previously examined studies to this research, cost asymmetric behavior is expected to be found as well.
Accordingly, following Anderson et al. (2003), the first hypothesis is:
H1: The relative magnitude of an increase in SG&A costs for an
2.2. Cost Stickiness and Managerial Influences
The traditional cost and management accounting theory not only implies a linear relationship between levels of activities and costs, but also assumes cost behavior to be independent of the managers’ decision regarding the resource-adjustment process and the handling of committed resources (Guenther et al., 2014).
However, the influence of managerial behavior on cost stickiness has been a topic of research for almost three decades (Cooper and Kaplan, 1988). The underlying assumption in this regard is that certain managerial behavior impacts the cost behavior.
Malagoli (1985) states that the omitted, premature, delayed, impossible, or inappropriate adjustment of the committed resources to changes in the level of the firm’s activity are the main reasons for cost stickiness. Managerial behavior further impacts a firm’s cost structure, for example, in the form of contract-entering for resources that are costly to break or to renegotiate. In such a case, if the firm faces a decline in demand, managers might prefer to retain underutilized resources rather than provoke high costs related to contract-breakage. Hence, while the firm might encounter a decrease in revenues, costs will not decline in the same amount as the revenues have fallen (Calleja et al., 2006). The term adjustment costs summarizes several forms of costs, such as equipment, severance payments and search and training costs for employment purposes. Next to this, adjustment costs also include indirect or psychological organizational costs such as a loss of morale among remaining employees in case of redundancy (Anderson et al., 2003).
Further, next to adjustment costs causing asymmetric cost behavior, Anderson et al. (2003) address (deliberate) managerial decision-making regarding adjustment costs and personal considerations. The latter partly depends on the manager’s personal characteristics as well as expectations in light of the near future. More precisely, the manager faces a decision to either remove or maintain committed resources. This decision depends on the personal assessment regarding a temporary or a long-term decrease in demand. In the first case, the manager might decide to maintain committed resources, leaving them unutilized for the period of low
As demand is stochastic, managers ought to evaluate the likelihood of a temporary or permanent decrease in demand when deciding whether to adjust committed resources correspondingly (Anderson et al., 2003). These decisions are based on several factors. Previously mentioned personal characteristics include working experience, risk-aversion and personal need for achievement (Papadakis et al., 1998).
Thereupon it is fair to conclude that asymmetric cost behavior is to be prevented only to a limited extend through a manager’s action, as adjustment costs naturally occur to a certain extent. Nonetheless, cost stickiness lies within managers’ area of responsibility as far as the degree of asymmetric cost behavior can be shaped by the managers’ decisions.
Cost behavior is a relevant element in managerial theory as it is an important aspect of profit analysis for managers: a firm that shows a high asymmetry in cost behavior shows a greater decline in earnings when the volume of the operational activity declines than a firm with a lower asymmetry in cost behavior (Weiss, 2010).
Anderson et al. (2003) and Anderson et al. (2001) argue that the phenomenon of cost stickiness finds its origin in asymmetrical responses of managers. That is, if the demand exceeds the expectations and the firm operates close to the maximum capacity, business resources are strained and managers are inclined to raise committed resources to satisfy the demand. Following, if demand falls behind expectations, slack resources exist which may be slow to eliminate. The managerial decisions in this regard hence impact the degree of cost asymmetry.
This study will measure cost behavior by means of SG&A costs. SG&A costs are a relevant cost measure as it captures most of the overhead costs that incur in corporate offices. Because the sales volume drives many components of SG&A costs, its behavior and the relation can be meaningfully studied (Cooper & Kaplan, 1998). By the same token, Guenther et al. (2014) stress the importance of SG&A costs, as they make up a large amount (26.4 percent) of sales revenue for their sample of 7.629 firms over a 20-year period.
Additionally, managers who are prone towards empire building tend to inflate SG&A costs too rapidly when sales go up or analogously decrease SG&A costs too slowly, respectively (Chen et al., 2012). Furthermore, Anderson et al. (2003) state that asymmetric cost behavior of SG&A costs occurs when managers decide to retain unutilized resources rather than incur adjustment costs when volume decreases. The literature on sticky costs predominantly studies short-term responses of SG&A costs to changes in revenue, focusing on managerial discretion regarding cost management in the current period (Balakrishnan et al., 2014).
2.3. Opportunistic Managerial Behavior
Opportunistic managerial behavior lies in the center of the agency theory. The central concern of the theory is solving problems and challenges that may occur in an agency relationship due to unaligned goals, different levels of risk aversion or related issues that origin in these predicaments (Jensen and Meckling, 1976). Other factors such as cognitive deficits, the fear of losing the job, adherence to old habits and sluggishness and lack of decision-making competence can further lead to inefficient behavior and further contribute to asymmetrical cost behavior (Mahlendorf, 2009; Malagoli, 1985). Further, managers might be reluctant towards downsizing, which in turn can contribute to asymmetric cost behavior, for three reasons. First, managers commonly benefit monetarily as well as non-monetarily from larger and more complex organizations, respectively organizational structures. Second, the
beneficiaries of downsizing tend to be the firm’s shareholders rather its managers. Third, managers might prefer to avoid painful and unpleasing decisions associated with downsizing for interpersonal reasons (Chen et al., 2012).
2.3.1. Empire Building
A central issue related to managerial behavior addressed by the agency theory is empire building. It has been a long recognized topic of research, as already Schumpeter (1934) posited managers to be empire builders.
Empire building describes the act of growing the firm beyond its optimal size or to maintain unutilized resources to increase the size and scope of an individual’s or an organization’s influence or power (Jensen, 1986; Masulis et al., 2016; Stulz, 1990). This is a stern issue as a firm’s growth has developed to form a common basis for executives’ monitoring,
performance evaluation and compensation (Hope and Thomas, 2008). A manager henceforth, consciously or unconsciously, might have personal motivations or monetary incentives to engage in empire building and increase the organization’s size or diversity operations above an optimal level.
Besides compensation and performance evaluation, scholars address several other possible motivations behind empire building: hubris (Seth et al., 2000), and executives’ desire for status, power, and prestige (Borges et al., 2014; Hope and Thomas, 2008; Jensen, 1986).
An illustration and an influential example of empire building is given in Jensen’s (1986) Free Cash Flow Hypothesis. The theory states that high levels of free cash flows incentivize managers to rather invest in projects, even if they show a negative net present value, than paying out the free cash flow to shareholders, and has been empirically proved in several studies (Gibbs, 1993; Lang et al., 1991; Richardson, 2006).
Empire building has become one of the fundamental issues in the literature on corporate governance. The underlying assumption is that managers turn into empire builders if not appropriately governed (Dominguez-Martinez et al., 2006). Further, Jiraporn et al. (2006) describe a manager’s lack of accountability towards the shareholders to increase the
probability of the manager engaging in empire building. Correspondingly, means to increase managers’ accountability should lead to less empire building, better firm performance, and, consequently, higher shareholder value. Bridging the gap to this research, the just mentioned benefits of increased accountability are expected to decrease the costs’ asymmetric behavior. Appropriately, Chen et al. (2012) described a positive association between incentives leading to managers’ empire building and the degree of cost asymmetry as well as a negative
Concluding on these findings: empire building leads to more pronounced cost stickiness. In turn, this relationship can be alleviated by strengthened corporate governance mechanisms. This conclusion provides this work’s hypotheses with supplementary substance.
All in all, empire building literature indicates several possible managerial motivations to engage in empire building. Empire building can result in suboptimal firm management, consequently leading to asymmetric cost behavior (Chen et al., 2012). The literature
furthermore suggests that empire building and its consequences can be mitigated by increased managerial accountability in the form of corporate governance.
Properly, corporate governance and its ramification towards cost stickiness will be examined in the following section.
2.4. Corporate Governance
Corporate governance is an umbrella term that covers several concepts, theories and practices of boards of directors, their executive and non-executive directors (Cochran and Wartwick, 1988). While academics previously have researched the influence of corporate governance on firm performance and managerial decision making (Larcker et al., 2007), Shleifer and Vishny (1997) described stronger corporate governance mechanisms to mitigate agency problems. This mitigating effect is empirically backed in regard of the managers’ tendency to engage in empire building: stronger corporate governance mechanisms lead to less pronounced
managerial engagement in empire building (Gaspar et al., 2005; Masulis et al., 2007; Titman et al., 2004).
Further, Chen et al. (2012) empirically prove that stronger corporate governance mechanisms can reduce the level of cost stickiness. The scholars find “strong evidence that cost
asymmetry is positively associated with managers’ empire building incentives due to the agency problem” (page 278) and moreover describe the relationship between agency
Moreover, the impact and benefits of these corporate governance mechanisms – economic and legal institutions that can be influenced by political processes – can be seen in less developed countries. In Russia, for example, the poorly developed corporate governance mechanisms lead to a considerable diversion of firm assets by managers of many privatized companies (Boycko et al., 1993). This is also in line with the aforementioned regarding opportunistic managerial behavior and empire building. In the context of this research corporate governance will be examined regarding its function to increase the manager’s accountability as well as improving the decision-making process with reference to cost behavior. As this study researches the impact of board diversity on the quality of corporate governance mechanisms, measured by the degree of asymmetric cost behavior, using a sample of US firms, the Anglo-Saxon corporate governance model and its implications towards cost behavior will be laid out in more detail in the following.
2.3.1. The Anglo-Saxon Corporate Governance Model
Main examples for the Anglo-Saxon corporate governance model are the United States of America, Australia, Canada and the United Kingdom. In this common-law system of
corporate governance, executive directors and non-executive directors form a one-tier board and operate in this organizational layer together (Maassen, 1999). Its implications towards cost behavior mainly find the origin in the model’s shareholder-orientation and its regulatory framework. Shareholder-orientation emphasizes shareholder maximization as a dictum, directing a cardinal role to the stock markets (O’Sullivan, 2003). To that effect, the stock market functions as a mechanism through which corporate control is exercised by the market, consequently rewarding (disciplining) overperforming (underperforming) management (Calleja et al., 2006).
These features indicate the management’s scope of power to be determined by a smaller, yet potentially more goal-congruent community of interests. Continuing, the emphasis on
shareholder-orientation increases managerial discretion in a way that enables management to take decisions that might dissatisfy certain stakeholders while appeasing shareholders
Concluding on the finding of the aforementioned sections in regard to this study, empire building increases the probability of a higher degree of asymmetric cost behavior.
Furthermore, as Chen et al. (2012) stated, corporate governance is directed and aiming at mitigating the agency problem, and, at the same token, empire building. According to the afore-cited scholars, this impact on a firm’s cost behavior is expected to be asymmetry-reducing.
As previously elaborated on, the effect of increased board diversity on corporate governance and its quality, in this study, is examined via the degree of a firm’s symmetry of cost
behavior. Meaningful studies reiterate and highlight corporate governance’s direct influence on managerial accountability, empire building, and, in consequence, on a firm’s cost
behavior. Another substantial idea of academic literature is directed towards diversity improving the functioning abilities of corporate governance.
Appropriately, relevant literature in regard to board diversity and its implications towards corporate governance will be surveyed next.
2.5. Board Diversity
Due to an increasingly international and thus complex business environment, as well as corporate scandals in the past, corporate governance and ways to improve its governance ability receive an increasing amount of attention (Agrawal and Chadha, 2005). As ways to improve the mechanisms of corporate governance are looked for, much consideration is paid towards the board’s characteristics and composition, such as its diversity (Mishra and
Jhunjhunwala, 2013b). Because the board’s decisions not only determine a firm’s success and long-term survival but further impact shareholders, employees, customers and all other stakeholders, the board must optimally fathom the firm’s direct and indirect environment, political, economic, and industry conditions and market developments (Gupta and Raman, 2014).
Accordingly, the question to be answered addresses the optimal composition and
Scholars designate a vital role in determining corporate performance to the board and its composition (Guest, 2009; Hermalin and Weisbach, 2001). All in all the results are partially ambiguous, but research suggests diversity to positively effect board performance, as diverse board members bring a broader set of unique perspectives to the boards (Arfken et al., 2004; Walt et al., 2006).
Previous research gives reason to further examine board diversity along the measures age and gender. Gender diversity not only is a hot topic in economics as well as politics, but further is presumed to improve board functioning (Hillman, 2015). At the same time, a board composed of members of varying maturity might improve its governance function as older board
members bring experience and wisdom to the board, while young members are inclined to be more energetic and affine towards latest technology and thus contribute to a board’s
governing quality through a more modern approach of governing (Carter et al., 2003). These findings, both individually and in aggregate, do not lead to a sufficient and reliable conclusion about board diversity and its impact on corporate governance themselves, but rather supply the hypotheses with further substance.
Diversity along the dimensions age and gender will be examined in the following sections.
2.5.1. Age Diversity
While age diversity to some extend naturally occurs at any board, the majority of board members commonly are of middle to senior age. This is due to the fact that board members usually show a similar pathway in terms of education, experience, and maturity (Kang et al., 2007). In theory, high age differences within a board automatically leads to a broader knowledge basis of the board of directors (Hagendorff and Keasey, 2012), which is due to natural differences in education, views and individual environmental influences.
Accordingly, to satisfy the potentially vast variety of stakeholders, a firm is properly governed by a board that contains members with corresponding multiple views and
perspectives (Huse and Rindova, 2001). Consequently, the resulting plurality of perspectives bears the potential to lead to improved decision-making and therefore improved corporate governance.
At the same time, other studies contradict these findings. Randøy et al. (2006) proclaim age diversity to not have a significant effect on stock market performance as well as return on assets of the 500 largest companies from Denmark, Sweden and Norway. Hagendorff and Keasey (2012) empirically find age diversity to contribute to wealth losses surrounding acquisition announcements for research in the US banking industry.
All in all, previous research shows ambiguous results respecting age diversity’s effect on company performance. Though, as age diversity permits the board to apply a broader set of perspectives and hence should lead to better decision making, corporate governance
mechanisms should be positively impacted as well. Considering the above, stronger corporate governance mechanisms could furthermore influence cost stickiness, which leads to the second hypothesis:
H2: Higher age diversity within the board leads to less pronounced asymmetric cost behavior.
2.5.2. Gender Diversity
Gender diversity within the board of directors and its (societal) development is a controversial topic. As a result of the debate, several European countries, such as Germany, Spain, France and Norway, have introduced legal quotas to emphasize and support board gender diversity (Ahern and Dittmar, 2012; Terjesen et al., 2014). To examine the impact of a higher share of female board members on firm performance, several studies with partly ambiguous results were conducted.
According to Hillman (2015), women do not only emphasize ethical practices more than men but also spend more time considering decisions, thereby potentially leading to improved decision making. As women represent half of the global workforce, academics observed that employees are more likely to stay within a firm when the firm’s board represents the
workforces’ demographics. To corroborate this matter, potential female and minority
employees see a firm with the aforementioned attribute regarding its board composition as a more attractive employer (Daily and Schwenk, 1996; Hillman et al., 2007). Furthermore, the presence of female board members enriches board quality as far as females more frequently attend meetings, indirectly stimulating male board members to do so as well (Adams and Ferreira, 2009). Also, Adams & Ferreira (2009) stress the importance of gender diversity in companies with weak governance arrangements. These findings indicate that the relationship between gender diversity and firm performance is influenced by the current governance arrangements. Ensuing, these findings suggest females to enrich a board’s quality regarding its function to govern a corporation.
On the other hand, doubts concerning the aforementioned findings exist. Carter et al. (2010), for example, find no positive empirical effects of gender diversity on firm performance. Other researchers state that the females’ limited experience regarding leadership and their lesser drive to proceed to the top hampers their effectiveness as board members (Dargnies, 2012).
Overall, the literature on gender diversity implies that the board composition containing male as well as female board members improves corporate governance. As improved corporate governance influences cost stickiness, gender diversity is expected to reduce asymmetric cost behavior. This leads to the third hypothesis:
H3: Higher gender diversity within the board leads to less pronounced asymmetric cost behavior.
All in all, the previously reviewed literature suggests an improvement in corporate
2.6. Degree of Internationality
In the last decades, internationalization has become a common option to secure organizational growth once demands in the local markets are satisfied (Al-Rodhan and Stoudmann, 2006). It enables firms to secure competitive advantages over purely domestic firms but at the same time is related to increased managerial, operational and organizational complexity. This complexity can be allegorized as a greater diversity regarding cultures, customers,
competitors as well as rules and regulations. Accordingly, internationalization is associated with great opportunities as well as challenges and threats, emphasizing possible risks and costs (Chen, 2011).
These circumstances lead to a greater demand towards information processing, coordination and agency requirements (Oxelheim & Wihlborg, 2008). Moreover, researchers emphasize resulting coordination difficulties, information asymmetries and increased monitoring difficulties and costs (Tihanyi et al., 2000; Zaheer, 1995).
The aforementioned leads to decisive implications concerning the composition of the board of directors. Scholars postulate firms to match managers to organizational strategies (Greve et al., 2009): in order to appropriately take on the previously stated challenges, a firm that engages in an internationalization strategy should be governed by a internationally composed board of directors. This diversity brings a wider range of expertise and experience regarding different institutional settings. Therefore, problem-solving as well as innovative thinking is improved and increased (Nielsen and Nielsen, 2012). Further, as strategic decision making is a highly complex task characterized by uncertainty and lack of routines, an internationally diverse board is, according to Nielsen and Nielsen (2012), likely to improve the process. This impacts firm performance and, in turn, influences its cost behavior.
McLeod et al. (1996) empirically show for small groups that ethnic diversity leads to more effective solutions as well as solutions of higher quality. Nielsen & Nielsen (2008) state and empirically prove for Swiss listed firms representing 32 industries that an increase in
Yet, Randøy et al. (2006) for example find that nationality diversity does not affect stock market performance and return on assets. In a broader sense, scholars concluded that,
regarding communication and interaction styles, internationally diverse teams may encounter conflicts, lower cohesiveness, and slower decision-making (Earley and Mosakowski, 2000; Hambrick et al., 1998).
Withal, although providing equivocal conclusions, the literature on internationalization suggests an internationally active firm to be governed by a nationality diverse set of board of directors. In turn, higher nationality diversity is likely to lead to better decision making of a board of directors, which subsequently is assumed to lead to better corporate governance and as such to lower asymmetric cost behavior.
Accordingly, the fourth hypothesis is:
3. Methodology
This chapter will first explain the regression analysis, followed by the definition of the dependent and independent variables, as well as the control variables and their measurement. Next, a description of how the data for this research has been collected is given.
3.1. Model
This study will research cost stickiness as the dependent variable applying linear regression analysis. The following basis model tests for average cost stickiness behavior with respect to hypotheses 1 and as such is in accordance with influential literature in cost behavior research (Anderson et al., 2003; Balakrishnan and Gruca, 2008; Calleja et al., 2006; Chen et al., 2012; Subramaniam and Weidenmier, 2003).
(1)
SG&A and Rev stands for selling, general and administrative costs and sales revenue, respectively, for firm i in year t. The error term is represented by 𝜀𝑖,𝑡. DD stands for decrease dummy, which is a binary variable that takes the value 0 when revenue increases and 1 when revenue decreases. This leads to 𝛽1 measuring the cost increase in percentage terms with a 1 percent increase in revenue. If the binary variable takes the value 1, the sum of 𝛽1 and 𝛽2
measures and shows the percentage change in SG&A costs when revenue decreases.
Consequently, if the traditional model of cost behavior holds, upward and downward changes in costs will be equivalent and 𝛽2 is 0. On the other hand, if cost behavior shows stickiness, the variation of the SG&A costs should be greater with revenue increases than the variation for revenue decreases. Accordingly, the sum of the coefficients 𝛽1 + 𝛽2 measures the
percentage decrease in SG&A costs when revenues decrease with 1 percent (Anderson et al., 2003). Continuing, a significantly positive 𝛽1 coefficient and a significantly negative 𝛽2
coefficient would imply asymmetric cost behavior, when the variation of SG&A costs with revenue increases is greater than the variation for revenue decreases. With respect to hypothesis 1, the results show cost stickiness when 𝛽2 < 0 under the condition of 𝛽1 > 0.
Based on the literature review and first hypothesis, this research expects the combined sign of 𝛽1 and 𝛽2 to be positive and statistically significant, evidencing cost stickiness.
The use of the log model ensures normality, and, in combination with the ratio form, improves the comparability of variables across firms and consequently simplify economic interpretation of the estimated coefficients (Banker and Byzalov, 2014). This is of particular importance to this research as the sample firms differ regarding e.g. performance, size or the control variables asset intensity and employee intensity. Furthermore the use of a log model is in accordance with previous studies and moreover reduces the potential for heteroscedasticity (Anderson et al., 2003; de Medeiros and Costa, 2004). The decrease dummy is left out as an separate variable as its impact is expected to be zero, and to improve the fit of the model.
In order to test hypotheses 2 to 4, the main model is extended with the respective variables and presented as the full model in the following:
(2)
This model extends the basic model by adding the diversity dimensions age and gender, AgeDiv and GenDiv, respectively. In order to test the hypotheses, the model will be expanded by the respective variables in a stepwise approach, until the full model is applied.
The presented model further is elongated by a variable measuring the degree of
internationality of firm i in year t as well as the control variables asset intensity and employee intensity, as further elaborated on in the following sections. The interaction terms will
examine the moderating effect of the respective variables on the relationship of revenues and costs in regard to the aforestated hypotheses.
In light of the aforestated hypotheses 2 and 3, an increase in the board of directors’ diversity along the dimensions age (𝛽3) and gender (𝛽4) should lead to an increased quality of the
board, which is expected to be expressed via positive coefficients with reference to these variables.
Likewise, in order to support hypothesis 4, the coefficient addressing a firm’s degree of internationality, 𝛽5, is expected to be significantly positive, denoting a reduction of asymmetric cost behavior.
3.2. Dependent Variable
The dependent variable in this model, in accordance to Anderson et al. (2003) and Chen et al. (2012), is SG&A costs. Other cost measures will be used in order to increase the robustness of the model’s results. These will be costs of goods sold (COGS) and operating expenses (OPEX), which is in line with the regarding literature (Subramaniam and Weidenmier, 2003; Yasukata, 2011).
3.3. Independent Variables
Several characteristics of the board of directors as well as the firms’ degree of internationality will be used as independent variables. In accordance with the aforestated hypotheses, these board characteristics are the diversity dimensions age and gender.
Following Hagendorff & Keasey (2012), age diversity will be measured using the standard deviation of the age of the members of the board of directors.
Gender diversity, namely the share of female board members, will be measured using the percentage of female board members as in Adams and Ferreira (2009), Carter et al. (2007) and Erhardt et al. (2003).
3.4. Control Variables
Several other factors can influence a firm’s cost behavior. These have been identified as employee intensity and capital or asset intensity (Anderson et al., 2003). Employee intensity characterizes the ratio of the total number of employees to sales, capital or asset intensity depicts the ratio of total assets to sales. Firms with high employee intensity face higher adjustment costs due to relatively high costs that accompany changes in personnel such as restructuring or redundancy (Guenther et al., 2014).
In similar vein, as assets have higher adjustment costs than services or materials, firms with high asset intensity are prone to asymmetric cost behavior (Anderson et al., 2003;
Subramaniam and Weidenmier, 2003). Both employee as well as capital intensity hence are seen to impact a firm’s SG&A costs, leading to cost stickiness.
Chen et al. (2012), on the other hand, found a negative relationship between employee intensity and the stickiness of SG&A costs. The scholars conclude the differing results to be caused by the difference in the sample selection. While both studies are cross-sectional, they research different time frames. Anderson et al. (2003) examine the time period from 1979-1998 and Chen et al. (2012) research a much younger sample from 1996-2005. Chen et al. (2012) assume this difference to be caused by the use of temporary labor. Accordingly, the firms in the younger sample might have taken advantage of temporary labor as it introduces a greater extend of flexibility in regard to employment.
This flexibility in form of the ability to cut employees when demand decreases and to hire employees on a relatively short-term when demand increases hence implies lower adjustment costs for temporary workers in comparison to permanent workers (Guenther et al., 2014), which in turn could lead to less asymmetric cost behavior.
3.5. Data Collection
In order to investigate the influence of diversity characteristics of the board of directors on a firm’s cost behavior, a sample of firms listed on the New York Stock Exchange is selected using data from the Orbis database. In accordance with Weiss (2010), the sample will contain only industrial firms with the SIC codes 2000-3999 (manufacturing division). This selection is based on two reasons. First, compared to utilities or other more regulated markets,
industrial firms generally operate in a more competitive market. Because managerial
discretion is better observable in competitive markets, industrial firms suit the purpose of this research best.
Second, a dataset comprising only industrial firms increases comparability of the results. By only including firms that are listed at the New York Stock Exchange, the availability of financial statements is ensured. Further, this circumstance avoids currency translation problems, hence increasing the comparability of the results of the study (Chen et al., 2012). The aforementioned selection hence provides the research with a diverse yet comparable set of industrial firms representing the world’s leading economy in gross domestic product (last updated: October 2016, The World Bank, 2015). This results in a preliminary sample of 528 firms.
Next, the firm sample is used to generate data regarding the diversity dimensions age and gender through MSCI GMI Ratings. The availability of the databases Orbis and MSCI GMI Ratings allow to test for a time horizon of the period from 2012 to 2015.
Relevant cost and revenue measures are obtained via the DataStream database. Further, the DataStream database provides this study with relevant data in respect to the variable
3.6. Descriptive Statistics
The following table shows the descriptive statistics:
Table 1:
Descriptive statistics for dependent and independent variables
Mean Min. Max. Std. dev.
Dependent Variable
SG&A costs 0.050 -0.766 7.988 0.235
Costs of goods sold 0.020 -0.998 10.175 0.311
Operating expenses 0.034 -0.737 6.993 0.207
Independent Variables
Revenue 0.035 -0.997 34.535 0.291
Age (Standard deviation) 13.919 4.330 31.139 6.179
Female (percentage) 0.100 0.000 0,500 0.072
Internationality 0.352 0.000 1.000 0.288
Asset Intensity 2.401 0.303 294.180 10.417
Employee Intensity 0.003 0.000 0.090 0.004
In order to ease interpretation, the logarithms have been transformed to percentage changes between two consecutive years. SG&A costs show the highest mean variation of the three cost measures with a yearly increase of 5 percent. Costs of goods sold (COGS) and operating expenses (OPEX) show a yearly mean increase of 2 percent and 3.45 percent, respectively. While the COGS show the lowest mean, they depict the highest decrease and increase. The highest decrease in COGS for the selected time period was 99.8 percent, whilst the highest increase was 1017.5 percent. These rather extreme figures have been manually backed in the dataset.
This is comparable to the average presence of females on corporate boards of directors (10.3 percent) of 67 countries, as Terjesen et al. (2014) note. Whereas the sample-firms concur with the average, they show similar numbers compared to Poland or Estonia.
The quotas, as introduced by countries such as Norway, Spain, France, Italy, or Belgium, range between 33 percent and 50 percent, which implies that the results, as seen in the descriptive statistics, although only representing one industry, are considerably lower than what the aforementioned countries consider as appropriate.
4. Results
In order to examine the effect of a firm’s degree of internationality as well as the board of director’s diversity along the dimensions age and gender on corporate cost behavior, linear regression is applied. Before conducting the linear regression, the Pearson correlation coefficient matrix is used to check the variables for multicollinearity. The results show no multicollinearity. Further, the results of the Hausman test, which was performed to test whether errors are correlated with regressors, suggest to use a fixed-effect model.
As stated in the previous section, the following model will be applied to test for general cost stickiness for the sample:
(1)
To control for considerable differences between the sample’s firms, logarithms are used. This further enables the results to be interpreted as percentage changes in costs and revenues. The dummy leads the model to distinguish between periods with increasing and decreasing sales revenues. When revenues increase, the dummy is 0 and 𝛽1 measures the cost increase in percentage terms when revenues increase with 1 percent. In contrast, when revenues decrease, the dummy takes the value 1, and, consequently, the combined coefficients 𝛽1 and 𝛽2 show the percentage cost decrease for the case that revenues decrease with 1 percent. If the traditional cost theory holds, 𝛽2 will be 0. In order for cost stickiness to be found, 𝛽2 will be
Hypothesis 1 tests for general cost stickiness for the sample firms. The results are presented in the following table:
Table 2:
Results of the linear regression
SG&A COGS OPEX
Constant 0.053 *** 0.044 0.022 * Revenue (𝛽1) 0.036 0.605 *** 0.672 *** DD*Revenue (𝛽2) -0.053 0.531 ** 0.062 Adjusted R-squared 0.003 0.685 0.673 F-Value 3.485 3.494 *** 3.499 *** Number of firms 486 495 500 Number of observations 1,447 1,473 1,483
The presented table shows the results of the first model with the cost measures selling,
general and administrative (SG&A) costs, cost of goods sold (COGS) and operating expenses (OPEX), as dependent variable in the respective columns.
The coefficients show the expected direction for SG&A costs as the dependent variable, with a revenue increase of 1 percent leading to an increase in SG&A costs of 0.36%. Yet, as neither the 𝛽1-variable nor the 𝛽2-variable is of significance, a conclusion regarding cost behavior is not possible. The lack of significance of both variables leads to the conclusion that revenues do not have a strong impact on SG&A costs, thus contradicting some of the
literature’s findings on the linkage between revenues and cost stickiness. The low value of the adjusted R-squared declares and the insignificance of the F-value underlines the impossibility to conclude on the results. Hence, hypothesis 1 is not supported.
This finding contradicts not only this research’s expectations, but furthermore the findings of previously examined academic research, implying anti-stickiness.
Nonetheless, this phenomenon itself has been observed and empirically shown before (Balakrishnan et al., 2004; Venieris et al., 2014). Venieris et al. (2014) conclude anti-stickiness to occur for firms with low organization capital, although for SG&A costs. For their research on cost behavior, Balakrishnan et al. (2004) examine the effects of the current capacity utilization on cost stickiness and suggest capacity utilization as moderating variable, as they find costs to exhibit greater stickiness when the firm operates closer to capacity thresholds. Thus, these results could be explained by the fact that the sample-firms do not operate close to that threshold. The scholars further state that cost stickiness rather occurs when a manager is optimistic about the future, although encountering declining business activity (Balakrishnan et al., 2004). Ceteris paribus, anti-sticky cost behavior, according to the scholars, can occur when management is pessimistic, which serves as an additional possible explanation for the results with respect to hypothesis 1.
Another possible explanation for the unexpected and reversed sign of the 𝛽2 coefficient could be given by Banker et al. (2006). Accordingly, sales changes in the prior year condition the outcome of model 1. Continuing, when sales decline in two consecutive years, the decrease in costs is larger than the increase in costs when sales increase: the expected results are reversed. Empirical estimation confirms that the expected signs of the coefficients, as formulated with respect to hypothesis 1, only occur when sales increased in the previous year (Banker et al., 2006).
The results for OPEX as dependent variable are similar to the ones of COGS, although lacking significance for the 𝛽2-variable. For both COGS and OPEX as dependent variables, the coefficient 𝛽1 (0.605 and 0.672, respectively) allows the following interpretation: COGS (OPEX) increased 0.605 percent (0.672 percent) per 1 percent increase in sales revenues.
Table 3:
Results Linear Regression with SG&A Costs as Dependent Variable
(1) (2) (3) (4) (5) (6) (7) Constant 0.045 0.055 0.055 ** -0.190 ** -0.201 ** -0.206 ** -0.196 * Revenue (β1) -0.313 0.109 0.076 * -0.115 0.025 0.210 -0.045 DD*Revenue (β2) 0.190 -0.218 -0.144 0.0259 -0.071 -0.161 -0.148 Interaction: Variable*DD*Revenue (β2) Age 0.023 -0.0006 0.002 Gender 2.021 0.069 0.318 Internationality 0.365 0.363 0.280 Standalone Variables Age 0.002 0.003 *** 0.003 *** 0.002 Gender -0.101 -0.061 -0.113 -0.035 Internationality 0.048 0.0475 0.072 0.050 Control Variables Asset Intensity 0.0274 0.0259 0.032 0.030 Employee Intensity 51.800 51.31 48.75 48.03 Adjusted R-Squared 0.003 0.003 0.003 0.013 0.012 0.013 0.011 F-Value 7.405 7.405 7.471 * 11.389 ** 11.389 ** 11.388 *** 17.388 ** Number of Firms 406 406 472 389 389 389 389 Number of Observations 1149 1149 1376 1076 1076 1076 1076
Similar to the results of model applied to test hypothesis 1, the adjusted R-squared shows low values, ranging from 0.003 to 0.026. Hence, these models are able to explain between 0.3 percent and 2.6 percent of the dependent variable.
Next, hypothesis 2:
H2: Higher age diversity within the board leads to less pronounced asymmetric cost behavior.
In order to test this hypothesis, the basic model is extended by the interaction variable addressing age diversity (model 1). No conclusion on this hypothesis is to be made, as the coefficients do not show any sign of statistical significance. Similarly, as the standalone variables as well as control variables are added to the model (model 4), the coefficients still do not show statistical significance. Henceforth, a conclusion regarding hypothesis 2 is not to be made.
Following, hypothesis 3:
H3: Higher gender diversity within the board leads to less pronounced asymmetric cost behavior.
Following, hypothesis 4:
H4: Firms with greater emphasis on international operations will show less pronounced asymmetric cost behavior.
Reiteratively, the basic model first is extended by the respective variable that aims at examining the moderating effect of a firm’s degree of internationality on cost behavior (model 3). Anew, the results would indicate that internationality reduces cost asymmetry, but the coefficients show no significance, except the 𝛽1-variable. This also is the case for the model that further adds the standalone variables and control variables (model 6). Encore, the result suggests internationality to reduce cost stickiness, but can not be interpreted in such way, as the results lack statistical significance. Accordingly, no conclusion can be made with respect to hypothesis 4.
Model 7 introduces the interaction variables addressing internationality as well as age and gender diversity. All three variables, as expected, show positive coefficients but show a dearth concerning their statistical significance.
To summarize, linear regression was applied with seven models and SG&A costs as the dependent variable. No conclusion can be made regarding the hypotheses due to a lack of statistical significance. Furthermore, the results, for the same reason, do now allow a conclusion regarding the control variables. Generally, low adjusted R-squared values imply that precise predictions or interpretations are not feasible.
5. Robustness Tests
As mentioned in section 3.4, this research uses other cost measures as robustness tests. These measures are cost of goods sold (COGS) and operating expenses (OPEX). Because the results only differ insubstantially between the two alternative cost measures, the results with COGS as the dependent variable will be presented and examined, as can be found in table 4. The results for OPEX as dependent variable are available upon request. Different from the
previous regression analysis and concurring with the cost measures’ results presented in table 2, table 4 shows adjusted R-squared values between 53.3 percent and 63.8 percent, indicating a much better fit of the model.
Similar to the linear regression applied with SG&A costs as dependent variable, in order to test for hypothesis 2 – the moderating effect of age diversity on cost behavior – the basic model is elongated by the respective variables in a stepwise approach. In model 1 both 𝛽1 and 𝛽2 coefficients show the expected direction, but only the 𝛽1 variable is significant.
Accordingly, a 1 percent increase in revenues leads to an increase in COGS of 0.824 percent. Still, when the model is added by the standalone and control variables, again, no conclusion regarding hypothesis 2 can be made.
To test hypothesis 3 – the moderating effect of gender diversity on cost behavior – an
interacting variable is added to the basic model first (model 2). The results show both the 𝛽1 and 𝛽2 coefficients to denote the expected properties and to be highly significant. Following, a 1 percent increase in revenue leads to a 1.149% increase in COGS, while a 1 percent
decrease in revenue leads to a 0.749% decrease in COGS (𝛽1 + 𝛽2), indicating cost stickiness and confirming the afore cited scholars (Subramaniam and Weidenmier, 2003; Yasukata, 2011). Further, the interaction variable of gender diversity becomes positive and statistically significant. Hence, higher gender diversity within the board of directors reduces cost
stickiness, hypothesis 3 is supported. This is in accordance with Gul et al. (2011).
Table 4:
Results Linear Regression with COGS as Dependent Variable
(1) (2) (3) (4) (5) (6) (7) Constant -0.007 -0.0009 0.086 ** -0.014 -0.040 -0.202 -0.029 Revenue (β1) 0.824 *** 1.149 *** 0.412 0.967 *** 1.113 *** 1.097 *** 1.074 *** DD*Revenue (β2) -0.038 -0.400 *** 0.807 ** -0.215 -0.420 *** -0.173 -0.380 Interaction: Variable*DD*Revenue (β2) Age -0.006 0.002 0.002 Gender 2.773 ** 2.241 * 2.475 ** Internationality -1.293 ** -0.0237 -0.091 Standalone Variables Age 0.0004 0.001 ** 0.001 ** 0.0006 Gender -0.167 0.018 -0.165 0.0396 Internationality -0.030 -0.029 -0.036 -0.042 Control Variables Asset Intensity -0.0261 -0.029 -0.021 -0.025 Employee Intensity 22.760 ** 24.530 *** 19.930 ** 21.950 ** Adjusted R-Squared 0.621 0.621 0.533 0.634 0.637 0.635 0.638 F-Value 7.411 *** 7.411 *** 7,480 *** 11.394 *** 11.394 *** 11.394 *** 17.394 *** Number of Firms 412 412 481 395 395 395 395 Number of Observations 1170 1170 1404 1097 1097 1097 1097
In accordance with the previous tests, to test hypothesis 4 – the moderating effect of internationality on cost behavior – the basic model first is expanded by the respective interaction variable (model 3). Contrary to the formulated expectations, a higher degree of firm internationality enhances cost stickiness, as the coefficient is significant but negative. Thus, hypothesis 4 finds no support and is rejected. These findings emphasize the
aforementioned risks of internationalization due to increased complexity and thereby confirm the respective scholars (Chen, 2011; Tihanyi et al., 2000; Zaheer, 1995).
With reference to the control variables it is just to conclude that employee intensity decreases cost stickiness, as the coefficients, opposing this research’s expectations, show positive signs and are statistically significant at the 5 level (model 4 and 6) and at the 1 percent-level (model 5).
These results contradict the findings of Guenther et al. (2014) as the scholars predicate employee intensity to lead to higher adjustment costs, thus contributing to cost stickiness. While contradicting the aforementioned, this research’s results confirm Chen et al. (2012), ascribing employee intensity to improve cost behavior, although for SG&A costs. Hence, this paper confirms a negative relationship between employee intensity and cost stickiness for cost of goods sold. A possible reason for this finding, as elaborated on previously, is the timeframe of this study. During the years 2012 until 2015, the sample-firms might have taken advantage of a flexible use of temporary labor.
6. Conclusion and Discussion
This paper researched the relationship between board of director’s diversity, firm
internationality and corporate cost behavior. The diversity dimensions age and gender were examined with respect to their improving effect on corporate governance functions, which in turn influences cost stickiness.
Previous literature on age and gender diversity in the board of directors find a more diverse board to increase the knowledge-base the board can judge, evaluate and monitor upon. These effects, in regard to this paper, result in a reduced asymmetry in the sample firms’ cost behavior. As less asymmetric cost behavior improves the firm’s flexibility with respect to changing economic conditions, it positively impacts firm performance.
Academic literature not only directs a pivotal role with respect to a firm’s performance to the board of directors, but moreover assigns the potential of improving corporate governance functions within the board to diversity factors. Thus, based on the assumption of opportunistic managerial behavior, grounded on theory regarding the agency theory and empire building, hypotheses have been formulated that postulate increased board diversity along the
dimensions age and gender to improve corporate cost behavior due to improved corporate governance functions. The hypotheses have been empirically analyzed with a sample consisting of 528 manufacturing firms, listed at the New York Stock Exchange, over the period 2012 until 2015. Analogously, a similar outcome regarding the impact on cost behavior was predicted with a higher degree of the firm’s internationality, as a more international firm is expected to incorporate a more diverse board.
The results of the linear regression with respect to hypothesis 1 disproved cost stickiness but rather proved anti-stickiness. Accordingly a 1 percent increase in revenues leads to an
increase in COGS of 0.605 percent, while a decrease in revenues of the same amount leads to a decrease in COGS of 1.136 percent. This abnormal behavior contradicts the findings of a large part of the literature (Anderson et al., 2003; Subramaniam and Weidenmier, 2003; Yasukata, 2011) and leads to the conclusion that hypothesis 1 is rejected.
Next, the effect of age diversity on cost behavior, with reference to hypothesis 2, has been examined. The results do not allow further interpretation, as the respective coefficients show no statistical significance.
Hypothesis 3 addressed the moderating effect of gender diversity, namely an increased share of female board members, on corporate cost behavior. The outcome of the regression analysis confirms not only cost stickiness to prevail, but also hypothesis 3: increased gender diversity, due to gender-based behavioral differences, reduces cost asymmetry, confirming Adams and Ferreira (2009).
Finally, in consideration of hypothesis 4, the influence of a firm’s degree of internationality on cost behavior has been examined. The results indicate internationality to increase cost stickiness. Appropriately, hypothesis 4 is rejected. This result might be caused by an increase in organizational and operational complexity that accompanies the process of
internationalization. Continuing, the results regarding hypothesis 4 must be interpreted with caution, as a high degree of internationality and thus organizational complexity might increase the firm’s reaction time and coordination demands with respect to adjusting corporate cost structures.
From an academic perspective, this paper contributes to the previous literature on a board’s composition and cost behavior. As major managerial implication, this paper proves that a higher share of women in the board improves the firm’s cost behavior. Moreover, this implies that higher gender diversity in the board of directors has beneficiaries beyond the firm, such as employees, customers and consequently society at large.
This is of particular importance for several reasons. First, this study improves the
Moreover, the results of this study can provide firm analysis with an additional perspective, as gender diversity has been proven to ameliorate corporate governance, which improves cost behavior. This can be of benefit for multiple parties, such as analysts, investors or
stockholders. As stated by Weiss (2010), the results of this paper can valuably contribute to more precise earnings forecast, which further enhances the potential of the aforementioned parties. Continuing, the rationale to hire female board members has implications beyond purely financial performance. Although supporting gender equality-enhancing laws and regulations, this research can contribute to a more general and societally vested approach: gender equality as an ethically and morally grounded principle, in this research exemplified via gender diversity within the board of directors, though implying these positive effects beyond the domain of corporate cost behavior.
Accordingly, it would be recommended to increase the share of female board members. This, in turn, gives further, and most importantly, empirically proven justification with respect to the introduced quotes regarding gender diversity in the board, as mentioned in section 2.5.2.
By virtue of a lack of statistical significance, the relationship between age diversity and cost behavior remains unclear. With respect to the firms’ degree of internationality, no universal or generalizable implications can be drawn for two reasons. First, the results do not allow precise interpretation. Second, circumstances accompanying corporate internationalization are too complex to reduce its impact on cost behavior alone (Sepasi and Hassani, 2015).This makes direct implications neither legitimate nor unambiguously possible.
This research’s findings and interpretation must be viewed in the context of the study’s limitations. Although the results do not lead to decisive implications with respect to age diversity, this does not suggest age diversity to not influence corporate cost behavior.
Further, this research only examined cost behavior for one industry (manufacturing) of firms, listed at one stock exchange (New York Stock Exchange). Continuing, only firms of
Internationality, as proxied by foreign sales over total sales, represents another limitation. As said, this variable has been measured using one out of many possible proxies, of which none fully grasp the content to an absolute degree.
Accordingly, future research should apply a dataset comprised of multiple countries in order to increase comparability. As mentioned in the section addressing the Anglo-Saxon Corporate Governance Model (2.3.1), the specific properties of this model might lead to different results for countries that apply a different model, as suggested by Calleja et al. (2006). Specifically, it is recommended for future research to include countries of both common law and code law systems.
