1
The influence of
Power on financial
performance
financialperformance
within Dutch Family
Businesses
Author
Marnix Teters 10022198
Supervisor
Dr. R. van der Voort
MSc Entrepreneurship Vrije Universiteit Faculty of Economics and Business Administration
Universiteit van Amsterdam Economics and Business
2
Abstract
Although family businesses have been an extensively researched topic within academics, there is still a lot of debate on how to define these organizations. Klein, Astrachan and Smyrnios (2002) have endeavored to provide handles to this discrepancy by developing an index that enables to measure the amount of familiness within a company as a degree instead of dichotomously categorizing businesses as family or not. Familiness is referred to as the unique characteristics a family business has over non-family businesses and will be explained in further detail within this paper. The F-PEC framework as this index is called stands for: Family influence on Power, Experience and Culture and is linked to financial performance of family businesses to discover possible relations. Klein, Astrachan & Smyrnios (2002) choose these constructs to make familiness measurable.
Within this paper the Power construct of the F-PEC index is examined. This construct measures share ownership and the influence the family has on governance and management. To understand why these items are important adjacent theories such as agency theory will be explained. The purpose of this paper is to examine if the degree of Power has a positive relationship with several financial performance indicators. The results show a positive sign though not significant. The other two constructs Experience and Culture which are the scope of two other papers using the same database make together with Power, familiness measurable. In the conclusion part of the paper it is addressed that an above average score on familiness is positively (significant) related to Net Profit Margin for the studies sample.
Keywords: Governance, Management, Succession, Power, Familiness, F-PEC, Financial Performance
The copyright rests with the author. The author is solely responsible for the content of the thesis, including mistakes. The university cannot be held liable for the content of the author’s thesis.
3
Table of contents
Abstract 2 1.0 Introduction 4 2.0 Theoretical Framework 9 2.1 Agency Theory 11 2.2 Stewardship Theory 13 2.3 Governance 162.4 Ownership and management structure 17
3.0 The Research Model 22
4.0 Methodology 24
4.1 Sample 24
4.2 Data Collection 24
4.3 Questionnaire Operationalization 25
4.4 Financial Indicators 27
4.5 Data Analytics Procedures 29
5.0 Results 30
5.1 Comparing sectors 35
6.0 Conclusion & Discussion 39
6.1 Limitations 44
6.2 Theoretical Contributions 46
6.3 Avenues for Future Research 47
7.0 References 48
4
1.0 Introduction
For the ordinary mortal, most important things in life are family and employment. The combination of these two has been an evolving field of inquiry over the last decades. However, due to the multidisciplinary character of a family business there is still much debate and heterogeneity to solve (Carney et al. 2015; O’Boyle et al., 2012; Anderson & Reeb, 2003). The heterogeneity and complexity of the subject could be a result of unknown moderators or mediators that have not been included in available papers yet (Mazzi, 2011).
Source: Harms (2014, p.291)
Harms (2014) analysed 267 journal articles to draw a picture of the academic landscape with regards to family involvement topics. As seen in the figure above, performance studies which mostly address financial (e.g. Return on Assets) and nonfinancial performance indicators (e.g. Corporate Social Responsibility) have been the predominant topic for family business research (i.e. 151 studies). The results of many of these studies are mixed (e.g. Anderson & Reeb, 2003; Martinez, Stohr & Quiroga, 2007; Sraer & Thesmar, 2007; Flören et al., 2010).
5
The definitional issue and search for clarification on what defines a ‘family businesses’ have been largely to the concern of the past, to a great deal of researchers in the respective research arena. Rutherford et al. (2008) refer to this as the ‘family business theory jungle’ to point out the lack of consensus because of many competing theories that examine familiness related to certain performance measures. Familiness can be interpreted as the bundles of resources and capabilities, that are distinct for a family (Bromiley & Rau, 2015).
Harms (2014) attempted to clarify the definitional issue by systematically clustering the relevant articles to obtain an overview of their definitional disparity.
Source: Harms (2014, p. 296)
As can be seen in the figure above, the majority of the articles in the author’s sample are without an explicit definition. Thereafter, a quarter of the articles consist of self-developed definitions. Accordingly, one can get a good intuition in how this missing unified definition concerning family business can cause problems in the research field. As Dyer (2006, p.254) concludes: ‘Thus, some
studies likely included firms in their ‘family firm’ sample that would not have been included in other studies’ samples and this mixing of ‘apples and oranges’ might account for the ambiguous findings’. Consequently, it does not come to a surprise that in the twenty-three studies Rutherford et al. (2008) reviewed between family involvement and firm performance, nine demonstrated
neutrality, nine studies showed a positive relationship, four studies provided partial support for a positive relationship and only one study found a negative relationship.
6
In general, family firms have advantages over ‘regular’ firms in the long term since it is understood that positive factors such as orientation, loyalty, teamwork, shared values and (reciprocal) altruism are more established within these firms; also known as the ‘bright side’ of family involvement. Contrarily, the ‘dark side’ poses possible downsides of family involvement such that businesses could be subject to conservatism, conflict, free-riding, deviance, and lack of professionalism (O’Boyle, Pollack, & Rutherford, 2012).
To overcome preceding definitional obstacles and measurement incongruities, the author of this paper adopted the F-PEC index. This index is created by Astrachan, Klein & Smyrnios (2002) and consists of three constructs namely: Power, Experience and Culture. Additionally, the F-PEC index has been validated and operationalized in the form of a questionnaire by Klein et al. (2005) and was further tested by Holt et al. (2010). The authors choose these constructs to capture familiness.
The F-PEC index is a scale to help understand the dynamics of family business on a continuous, rather than a dichotomous (i.e. simplistic categorization) scheme (Harms, 2014; Holt et al., 2010). This means that businesses are not merely defined as family or nonfamily, but that there is a certain degree of family involvement referred to as ‘familiness’. These three dimensions of involvement of the family within the business are measured according to this F-PEC index.
The abbreviation, F-PEC stands for: Family influence on (1) Power: the influence the family has on governance and management of the firm. (2) Experience: the information knowledge, judgement, and intuition that comes through successive generations. And (3) Culture: the alignment of the family’s goals and with the firm’s (Cliff & Jennings, 2005 p.342).
7
Although the F-PEC index has been an evolutionary step in family business research practical applications in the field are scarce. This study tries to apply the F-PEC index to Dutch family businesses and investigates if it is possible to find relations between high levels of Power and financial performance.
One of the few studies that tried to operationalize the F-PEC index is a study by Rutherford et al (2008). Unfortunately, they did not find evidence that Power has a positive effect on financial performance, they though did find stronger results for the other constructs: Experience & Culture although these were also mixed. Experience was positively related to revenue and the Culture construct was positively related to perceived financial performance. This paper will try to further examine if there can be found a relation between Power and financial performance within Dutch family businesses.
The figure below arrives at all the F-PEC index constructs with its corresponding dimensions. In the next chapter, the literature review starts which explains why it is interesting to address the Power construct.
8
The question raised whether the family specific attributes offset the disadvantages will be directed in the literature review part of the study. It could namely be the case that the family effects cancel out, suggesting neutrality of Power to performance. It could also be the case that the disadvantages dominate the advantages derived from Power of a firm. To come to a meaningful conclusion which of the three scenarios prevail understanding is needed of what defines Power within a family business. The research question proposed is: ‘Does the degree of Power have a positive influence
on financial performance?’
To say something meaningful about family businesses, it is important to know which factors distinguish a family business from a non-family business. Therefore, in the following chapters knowledge of adjacent theories (e.g. agency theory) for Power within the family business will be explained. These theories will provide a grasp of the underlying mechanisms that are at play within the Power construct. Stewardship and agency theory will provide handles to the fact why Power can be an asset or hurdle within the family business.
This research is initiated by Ernst & Young (EY), the Universtiteit van Amsterdam and the Vrije Universiteit. The family business department of EY asked the universities if they could perform a research to compare the financial performance of Dutch family businesses with public non-family firms. The results of this study can be found in a separate paper but the reason this is mentioned, is because the database used for that research is also used for this paper. Also, EY asked to make a distinction between sectors and this will be more deeply discussed in the methodology and results section.
A collective database has been created to assess the level of familiness which is a weighted score of the three constructs already mentioned. The three constructs of F-PEC are divided in three specific papers whereas Power, Experience and Culture have the central role in each of the theses.
As said, his paper focusses on the Power construct of the F-PEC index and in the discussion part the results of the combined findings of F-PEC are brought to the attention.
9
2.0 Theoretical Framework
As already briefly explained the links between family involvement and financial performance reveal very heterogenous findings (Rutherford et al., 2008). To assess if family involvement can result in better financial performance, Garcia-Castro & Aguilera (2014) have made a summary of more than 59 studies that tried to lay hands on the possible relationship. Most of the studies analyzed used four components to examine the relationship. Respectively: ownership, governance, management and succession. The first three components are measured within the Power construct of the F-PEC index and succession is addressed under the Experience construct of the F-PEC index.
The Power index consists of the family’s potential to influence the business through ownership, governance and management (Astrachan et al., 2002). In general, it attempts to assess the overall influence and the degree hereof by measuring the percentage of family members within three categories (i.e. ownership, governance and management). The conceptual model of the Power index consists as follows:
Source: Astrachan et al. 2002 (p. 47)
To understand the dynamics of these formations it is important to get an understanding of Agency & Stewardship theory which lay at the foundation of the mechanisms at play. Afterwards the dynamics of governance, ownership and management structures which are based on these
10
fundamental theories, within family businesses will be explained. The end of the Theoretical Framework will provide a recap and a hypothesis is formed.
11
2.1 Agency Theory
The classic principal-agent conflict described by Jensen & Mecklin (1976) could be one of the factors creating or destroying value within family businesses. Agency theory focuses on relationships between owners and managers which could result in a so-called type I agency problem. Within family businesses it would be possible that nonfamily managers make decisions based on self-interest that impact others (owners) in a negative way.
Private family firms can be understood as ‘pure’ ownership forms which could be less susceptible to principal-principal (example conflicts between family managers) agency problems (Carney et
al, 2015).
Whereas, secondarily, agency theory focuses on interactions between majority and minority stakeholders; a type II agency problem (Garcia-Castro & Aguilera, 2014)1. Typically, agency theory concerns the principal-agent problem, whereat the principal (e.g. the owner) delegates decision-making to the agent (e.g. the manager), where most commonly this results in misaligned interests. This is due to information asymmetries that can result in moral hazard. That is, the agent acts in favor of its own preferences that can result in opportunistic behavior (Schulze et al., 2001).
Especially, for dispersedly owned companies, type II agency problems occur. As Carney et al. (2015, p.519) provide a good example in that ‘[...] professional managers often enjoy substantial
freedom to pursue self-serving courses of action, which may include building large and diversified corporate “empires,” enjoying extensive managerial perquisites, ratcheting up their own pay and making it less contingent upon firm performance, and pursuing entrenchment even after documented poor job performance’.
One can imagine a family member (manager) enjoying the same freedom as the manager in the preceding description. Therefore, agency costs need to be incurred to align the interests (e.g. goals and visions) of the two to ensure that the two actors behave according to a shared strategy. This can be done by creating (complete) contracts that expand on eventualities by means of covenants to prevent opportunistic behaviour of the agent (Kraiczy, 2013).
12
Another example of a type II agency problem which appears in family businesses is that when large family shareholders tend to extract private benefits from the company which ultimately can negatively affect smaller family or nonfamily shareholders resulting in less financial performance. Claessens et al. (2002) explain that which of the two agency problems is more detrimental to shareholder value is still inconclusive. Although, family management has the potential to lower type I agency problems it would be possible that associated costs of family management would offset these advantages since it can be the case that hired professionals are better than the family founders or their offspring.
Carney et al. (2015) explain possible agency costs in greater depth. It is argued that the agency costs and benefits of private family firms lead to different strategies and decision-making processes compared to nonfamily businesses. For this reason, it is proposed to view private family firms as a distinct organizational form. One agency benefit of private family firms is that the high/pure ownership accentuates high-powered financial incentives. Another benefit of private family firms is that the absence of capital market pressures allows for a long-term orientation. Agency costs, on the other hand, are loss aversion, altruism and lack of moderation of noneconomic goals by the capital market pressures. The loss aversion for example, is explainable because the family often thanks their social status to the success and existence of the firm, and are willing to preserve this at all costs. These noneconomic goals can be harmful for the efficiency of the business (Gomez-Meija, Cruz et al., 2011). As already briefly explained, nepotism could also be also present in these private family firms because the relatives of the owners, which not always have the best competencies for the job could be favored over better candidates (De Massis et al. 2016).
The previously mentioned agency costs and benefits of private family firms are considered to be different from the agency costs and benefits of public family firms and both private/public nonfamily firms. These structures provide further reasons to argue that private family firms should be seen as separate types of business with different strategies and decision-making, which all have its own effect on firm performance (Carney et al, 2015)2.
2 See Appendix Figure 2 for an overview of costs and benefits based on agency theory for different organizational
13
2.2 Stewardship Theory
Stewardship theory can be considered as the other side of the coin (i.e. the ‘good side’). Consistent with stewardship theory, Mazzi (2011) explains that many business managers and leaders do not merely aspire only on goals that maximize their short-term, individualistic utility. Instead, stewardship is performed as they altruistically behave in a righteous path, with eyes to obtain beneficial results for the whole organization because stewardship can be formulated into competitive advantages. Especially, this seems to be prevalent in family firms where leaders are linked emotionally to the family. Stewardship can be recognized as the potential for the formation of superior performance and the creation of competitive advantages over companies that are subject to agency theoretical features (Eddleston et al., 2012).
Namely, long-run benefit for future generations, ensuring entrepreneurial longevity and continuity of the firm and its vision and mission. Moreover, it ensures unity and stability in the organizational structure through ‘patient capital’, which could be R&D, reputation-building and gradual acquisition of larger market shares. Secondly, stewards behave in different ways to employees and co-workers that foster a favourable working environment. This is conducive to the emergence of talented groups of people that is mediated through trust and loyalty that resides in the corporate culture. Thirdly, stewardship promotes long-lasting relationships where trust is central, with third parties such as shareholders (Mazzi, 2011).
Mazzi (2011) states that “noneconomic advantages and socio-emotional wealth refer to aspects of
the firm that are emotionally linked to the family’s affective dimension, such as protection of family ties, independence and continuity of family influence, perpetuation of the family dynasty, relationships with employees, social reputation and identity, links with the territory and the local community and so on” (p.176-177).
Typically, family firms have both noneconomic and economic goals. Therefore, Mazzi (2011) explains that the usual stream of wealth creation should be substituted with a more complex variable of performance, to capture the value creation aspects. This value creation resides in the capacity of family firms to leverage and develop assets that are more of intangible nature, such as trust, tacit knowledge, reputation and social capital.
14
Social capital could be interpreted as a part of the unique resources a family firm has in comparison to nonfamily firms. The family constantly uses, builds upon, and is by itself a spring of social capital. Namely, it can exist by trust, connections and relationships between different actors within the individual and/or organizational level. Kraiczy (2013) explains that social capital exists of cognitive, structural and relational dimensions and that value can be created through these complex structures. Sirmon & Hitt (2003) use a nice example to explain how social capital within the family business can be an asset: ‘[...] the family’s social capital increases by connecting these diverse
social structures, the firm can build more effective relationships with suppliers, customers, and support organizations (e.g., community financial institutions), while maintaining legitimacy with other important constituencies. In so doing, family firms garner resources from their constituencies and networks (e.g., knowledge, financial capital, and so forth). Additionally, they can more easily communicate the value of the firm’s goods and services to potential customers’ ,
(p.342)
Furthermore, future research ought to verify how family firms mitigate the ‘dark side’ of altruism and the entrenchment effects within which thresholds. Managerial entrenchment can occur if managers acquire so much power that it is possible to use firm resources at own interest instead of the interests of others or the firm.
For instance, Habbershon and Williams (1999) argue that these effects can be lowered due to a shared ‘family language’. Family member employees can be more productive if they are able to communicate better, wherewith private information can be exchanged more efficiently and effectively.
Building on this, family relationships result in higher motivation, trust and loyalties. As business strategies and family objectives are considered to be inseparable, it creates a better long-run strategy as well, as the commitment to accomplish it. For example, strong family communication/relationships can result in the existence of certain routines (efficient meetings) which can improve the level or professional management and dynamism, which then results in higher productivity and can ultimately result to higher profits.
15
Concluding, the two theories are relevant for this paper, as they are considered to be two opposing ‘forces’ that affect a firm negatively or positively. The ultimate question is which of the two is dominant in a firm or which one has the upper hand that ultimately affects firm performance in a (dis)advantageous manner (Eddleston et al., 2008). To get a better understanding of these interactions the next two chapters build further on these theories by investigating governance, management & ownership structures.
16
2.3 Governance
Miller & Le Breton-Miller (2006) explain how different governance structures that exist within family businesses can make family businesses great competitors. Based on previous research four distinctive dimensions of governance emerged, namely: (1) the family’s ownership & control structure, (2) family management involvement, (3) multiple family owners & managers, (4) involvement of multiple family generations.
These dimensions are part of the F-PEC index as well and need some further clarification to be fully understood. The above-named governance dimensions have the potential to create different choices and distinctive firm capabilities which can be positive or negative for the financial performance of the company (Miller & Le Breton-Miller, 2006).
Source: Miller & Le Breton-Miller, 2006 (p. 75)
The four dimensions are influenced by agency perspectives and can possibly result in distinctive capabilities. To recapitulate, from an agency point of view, it is understood that ownership concentration can reduce monitoring costs because significant shareholders very often have the knowledge and information to monitor and control their managers.
Anderson & Reeb (2003) explain that not only economic self-interest is what drives managers and owners but that these people also have the mission to make and keep the business and its stakeholders satisfied (Davis & Harveston, 2000). Family members have a strong emotional investment in the company which is acknowledged by Ward (2004) who explains the importance of personal satisfaction, public reputations which are at stake and the family's fortune which need to be protected.
17
Anderson & Reeb (2003) further explain that when members of the family serve as CEO in the family business the performance of the firm is better than with outsider CEOs. They make the notion that when members of the family serve as CEO agency problems can be reduced.
Further investigation of the amount of family ownership in a firm, revealed that the relationship between performance and family ownership is not positive in all levels, arguing that performance is first increasing and then decreasing with the level of ownership. Firm performance is increasing until families own about one-third of the firm’s outstanding equity. Beyond this level, performance starts to decline but is still better, on average, then in nonfamily firms. The study of Anderson & Reeb (2003) finds no evidence that the age of a family business has influence on financial performance but overall both young and older family firms tend to perform better than their nonfamily counterparts.
Gómez-Mejía, Nuñez-Nickel, & Gutiérrez (2001) argue the presence of competitive disadvantages in family businesses. These competitive disadvantages arise because family firms tend to favour family members when filling in managerial positions within the company. This is in agreement with Anderson & Reeb (2003), who argue that family firms often limit management positions to family members (nepotism). This restricted labour pool could lead to a competitive disadvantage compared to non-family firms. Villalonga & Amit (2006) explain that family management can add value when the founder serves as the CEO of the family business or as its Chairman with a nonfamily CEO, but destroys value when descendants serve as Chairman or CEO. Due to the previous issues, Morck, Strangeland & Yeung (2000) state that family ownership and control tend to result in poor firm performance. The academic landscape is thus still very mixed about this matter. In the following chapter, the ownership and family management structure is explained.
2.4 Ownership and management structure
The exploratory study of Naldi et al. (2015) investigates the relationship between family performance and family members serving in an advising capacity. Drawing on a sample of 128 Swedish family firms, they predict an inverted U-shaped relationship between family firm performance and the number of family advisors. According to Naldi et al. (2015) the definition follows as, ‘[...] family member advisors comprise those members of the owner family who are not
18
basis. Because they are often reluctant to provide nonfamily members access to information, family firms often rely on members of the owner family as advisors’ (p.4).
Similarly, the authors argue that family firms are emotionally charged and characterized by intense personal and professional interaction between family members. This casts no doubt on the decision-making process as having an effect through the emotional attachment and rational judgment, because they simultaneously invest in both their personal and professional lives. As aforementioned, family firms strive for a good community culture, and long-term continuity with strong relationships between staff and other stakeholders (through stewardship). Because stewards tend to place objectives of the firm ahead of their own, it aligns their motives with that of the business. This reduces conflict, fosters trust and strengthens family relations.
As explained, agency theory suggests that family members could be driven by self-interest, and use the business for perquisites. It is also possible that family members only serve the family and its needs, at the expense of other shareholders by underinvesting in the firm and extracting resources for personal purposes. Moreover, by excessively avoiding risk, they do not invest in projects that may renew the company in respect of product and process innovation. Therefore, the complementary theories (stewardship & agency theory) are at play, being positive or negative in the aggregate to financial performance of the focal firm. As Naldi et al. (2015, p.230) state: ‘We
expect that the performance benefits stemming from stewardship behaviour will be offset by monitoring and agency costs associated with high numbers of family advisors’. As a result, an
inverted U-shaped relationship is observed between the amount of family member advisors and firm performance, as seen in the figure below.
19
Source: Naldi et al., 2015 (p. 34)
Because, higher numbers of family member advisors, means a higher number of linkages of communication between them which adds to the complexity of coordination. This, in turn is detrimental to the effectiveness and efficiency of the communication and consequently, the potential negative outcomes hereof. Moreover, as the number of advisors increases, each of them may experience diminishing motivation as their contribution to the business is less identifiable.
Then, there is also a possibility of competition and conflict between the family member advisors. These are laden with sentiments and based upon emotion and informal linkages that result in less effective monitoring and disciplining of the advisors. The relationships in conflict can result in stewards becoming agents as result of frustration. Thus, at first hand, the arguments of stewardship favour firm performance with the increase of family member advisors. Though, across a certain threshold, monitoring costs, conflict and group think arise which result in a decline in family firm performance (Naldi et al., 2015).
The generation of the founder’s family that owns the firm serves as an important moderator. Because in the first generation, entrepreneurial spirit is high and risk of adopting risk averse behaviour is low. There is no intention to maintain a certain strategy of earlier generations in order to preserve wealth in the same way as the ancestors did. This strategy –that is emotionally attached to ancestral values– may have had contextual fit in that time, whilst it may have become dysfunctional in the changed context (i.e. the present), leading to a decline in performance.
20
Source: Naldi et al., 2015 (p. 35)
Similarly, De Massis et al. (2015) tested the relationship of family ownership in the top management team (TMT) and performance wherewith they also depart from the agency theoretical view. The authors complement the latter theory with behavioural theory, suggesting that the lack of self-control due to lower monitoring as a result of altruism (and also no public listing scrutiny), managers become risk averse. Herewith, having an impact on strategic decision-making that is unfavourable for the firm (De Massis et al., 2015). Similarly, the authors predict that family inclusion in ownership is expected to be favourable for performance, excessive family involvement in ownership impairs firm performance as a result of lack of scrutiny and control. Equivalently, they hypothesized –and found support for– an inverted u-shaped relationship between the performance (measured within the two respective years in terms of Return on Assets & Return on Equity)and the degree of family ownership. Indeed, the result shows a fitting resemblance of that with Naldi et al. (2015), with reference to the figure below.
21
Source: De Massis et al. 2015 (p. 936)
The findings of Naldi et al. (2015) & De Massis et al. (2015) are contingent with a group of scholars who argue that the effect of family involvement indeed has an inverted U-shaped relationship to firm performance (Anderson & Reeb, 2013), Dyer (2006), Mazzi (2011), Sciascia & Mazzola, 2008).
The mixed findings make it difficult to predict if the degree of Power within the family business will be positive or negative to firm performance. The costs of altruism, loss aversion and the pursuit of noneconomic goals which are unmoderated by capital markets can make the influence of the family a disadvantage for firm performance. On the other hand, pure ownership accentuates possibly high-powered financial incentives and the absence of capital markets pressure and the general stewardship principles can have a positive effect on firm performance as well.
Considering the three dimensions of Power, the abovementioned family factors are contingent on the composition of ownership, governance and management that can potentially unveil performance directions (i.e. positive or negative). Because of the disparity within the academic landscape it is even more interesting to unveil possible directions. Two hypotheses are formulated:
H0: The degree of Power has no influence on firm performance within Dutch businesses.
22
3.0 The Research Model
The information from the theoretical framework regarding the Power construct of the F-PEC index has enabled this study to test if there is a relation between Power and financial performance.
H0: The degree of Power has no influence on firm performance within Dutch businesses.
H1: The degree of Power has a positive influence on firm performance within Dutch businesses.
Source: own illustration
The Power construct is theorized as the independent variable and the financial performance indicators as the dependent. This was done to evaluate if directions and relations could be identified. A linear regression is performed for each of the performance indicators.
The findings for Power have been categorized in ‘high’ (above average score) and ‘low’ (below average score) and make it possible to identify differences between the two groups, this will be explained in greater detail in the methodology. The other constructs of the F-PEC index; Experience and Culture are the main subjects of two other papers that have used the same database as this research and that’s why the findings of these studies will need some clarification as well. In the discussion part of this paper the combination (familiness) of these constructs will be discussed.
Astrachan, Klein & Smyrnios (2002), could be regarded as ‘the founding fathers’ with respect to the inception of the F-PEC index. In order to give meaningfulness to the latter scale, the authors
23
and a team of skilled people, such as academic researchers, practitioners and family business owners developed the questions necessary to develop the F-PEC index in aggregate. Indeed, the authors proceeded with pilot testing and focus group discussions with a number of family businesses in order to validate the index with analysis of modelling techniques.
Evidently, items that were shown to demonstrate ambiguity, lack of discriminatory power or redundancy were eliminated (Astrachan et al., 2002). Furthermore, external validity was tested on a large sample of groups (i.e. n > 500) by means of cross-cultural comparison. As Astrachan et al, (2002, p. 52) conclude: ‘The F-PEC index of family influence on the business provides researchers,
for the first time, with a tested standardized instrument that allows integration of different theoretical positions as well as comparisons of different types of data’. Later on, Klein, Astrachan
& Smyrnios (2005), further tested the F-PEC index in a sample of 10.000 randomly selected company CEOs through exploratory and factor analytical techniques, to prove reliability (Klein et
al., 2005).
In addition, Holt, Rutherford & Kuratko (2010) further tested the measurement properties of the F-PEC index. Hence, further testing the items and making the operationalization more robust. Thereby, giving power to the generalizability, internal consistency and reliability of the different constructs (Power, Experience, Culture). Exploratory and confirmatory factor analysis were conducted and provided support to the index. As Holt et al., (2010, p. 84) discuss: ‘Although we
are not by any means suggesting that the F-PEC represents an end to the search for definitional clarity, our findings do suggest that the F-PEC offers family business researchers a valid and reliable scale with which to more finely classify family firms.’
24
4.0 Methodology
4.1 SampleThe sample used for testing the Power construct of the F-PEC index is selected by looking at the ‘Elsevier Top 100 familiebedrijven’ and the ‘FBned’ database. The Elsevier top 100 list of family businesses is a yearly recurring list and provides an overview of the top 100 largest Dutch family businesses when looking at the company's turnover. This study looks at the 2016 edition of the Elsevier Top 100 family business (Elsevier, 2016).
The final sample was constructed by selecting 93 companies with at least a turnover of 14 million. Larger family companies have been chosen for this research since the F-PEC calculation can be biased within smaller firms, which will be explained in more detail in chapter 4.3. According to Rutherford et al. (2008), there is also more robustness in the relationship between the F-PEC constructs to financial performance for larger family firms. Also, EY approved the companies which have been used in the sample.
Eventually 41 companies filled in the questionnaire, resulting in a response rate of 44%. However, three companies have been removed from the sample because there was not enough financial data available. Exclusion of these companies lead to a total of 38 companies which entirely filled in the questionnaire and had a complete the financial data set.
4.2 Data Collection
In order to find relations between the degree of Power and financial performance, information from a questionnaire and year reports have been collected. To every family business in the sample, a questionnaire has been sent in order to collect the information needed to assess the level of Power within the company. The questionnaire has been made in Qualtrics and sent by E-mail. The contact-details of each company have been obtained through a database called Company.Info, LinkedIn and Orbis. An Excel spreadsheet was made to organize the progress of the respondents. In this spreadsheet the name of the company, the industry, if the company is a client of EY, the name and function of some executive board members, phone numbers, email addresses, previous activity and the next activity was registered.
25
Before the questionnaires were sent, each company was contacted by phone in order to increase the response rate and to ask for email addresses to send the introduction mail together with the questionnaire to. The questionnaire was sent to participants of the founding family or members of the highest management or governance boards. Due to the extensive network of EY and the personal way of reaching out to the companies, the eventual number of responses was compelling.
4.3 Questionnaire Operationalization
The questions asked within the questionnaire have been obtained from previous F-PEC literature (Klein et al., 2005 & Holt et al., 2010). First, the respondent was asked whether he or she is part of the founding family. As a control question, it is asked whether there have been important changes in family involvement since the year of 2010. If that question was answered with ‘yes’, a follow-up question ought to be elucidated with what these important changes have been. If answered with ‘no’, the respondent is guided to the next question. The question: ‘If you had to
choose between the following sectors: food & drinks, Construction, Industry, Consumer Goods, which sector would best fit your company?’ was asked to evaluate in which of the four sectors the
particular companies do business. The industries are chosen in consultation with Ernst & Young and make it possible to evaluate if there are noticeable differences between the industries.
To assess the level of Power within a family firm three questions have been asked. The first question identified the proportion of share ownership by all family members (on a slide scale from 0% to 100%) within the firm. This question is obtained from Klein et al. (2005). Secondly, the ratio of family members was surveyed within the management and governance boards of the respective companies (Klein et al., 2005). The only difference is that within this research two separate questions have been asked to identify the total amount of members in a particular board, followed with a question to ask how many of these people where family members. The ratios have been calculated manually.
The three factors representing the Power construct are ownership, management and governance. Each of the factors within the Power construct accounted for ⅓ of the total Power score, since there are three questions asked to assess Power within the firm (see appendix for full questionnaire). The scores of these factors are expressed in percentages. For instance, if the management board consists of three family members and two non-family members, the score
26
attributed hereunto is 0.6, (i.e. 60%). If the company did not have a governance board the variables ‘ownership’ and ‘management’ are weighted as ½ each. One can understand that small family businesses overall have high scores on Power since mostly management boards within these smaller family businesses consists of few people and share ownership is high. That is partly why this study focusses on larger family businesses, the discussion part will provide more in-depth arguments for this decision.
To explain how the Power construct is operationalized in detail an example is provided: the management board consists of 6 people, where 2 are family. The ownership percentage of the company is 40%. The governance board consists of 10 people whereas only 1 is family. The calculation will look as follows: (1/3*0,25) + (1/3*0,40) + (1/3*0,10) = 0,2475. The sample of 38 companies all have a different score of Power. To make it possible to identify differences between companies it is chosen to search for the average score of Power from the sample. Companies with a lower than average score are categorized as ‘low degree of Power’ and those who scored above average as ‘high degree of Power’.
27
4.4 Financial Indicators
Financial ratios were gathered for the 38 companies. This resulted in 1.248 ratios and with these results 1.094 financial ratios could be calculated (not all the information was provided in the year reports). The six financial indicators used to measure financial performance were chosen in consultation with EY and are calculated as an average over six years. The data has been gathered through Orbis which is a database with financial information of companies and through Company.info which is a system with similar functionalities. Company.info was granted access to by EY. The discussion section of this paper will provide more details concerning possible shortcomings of these systems.
Revenue growth rate is measured as an average over 6 respective years.
Solvency, is equity divided by total assets in a percentage. This will show the ratio between equity and liabilities on the balance sheet. This gives an indication if the company is able to pay its debts (Heaton, 2007) and is also taken as an average over 6 years per company.
Net Profit Margin, is calculated by dividing the net profit with the revenue multiplied by 100%. This shows an indication of the profitability of the business (Blaine, 1994).The average of the Net Profit Margin is taken in consideration over the six respective years
Return on assets (ROA) is calculated as the net profit divided by total assets. It is a key figure indicating the profitability of the average total assets before interest deduction.
Return on equity (ROE) is calculated as the Net Profit divided by the Equity in a percentage. It is a key figure indicating the profitability of equity before interest deduction. This is like the ROA an accounting-based indicator to capture a firm’s internal efficiency (Cochran & Wood, 1984).
Return on capital (ROC) is calculated with the Net Profit, divided by the equity and the long-term debts. The ROC attempts to measure the return earned on capital invested in an investment (Damodaran, 2007).
28
To make the use of the ratios clearer an example is provided: company A has respectively 7%, 9%, 4%, 3%, 5%, 8% Return on Assets from 2010-2015 the average is 6% ROA for company A. This is done for all financial indicators. The amount of Power is measured on a fixed point in time and that’s why the decision is made to evaluate the averages of the financial indicators instead of the ratios of all the individual years.
29
4.5 Data Analytics Procedures
The data collected from the questionnaire and year reports have been processed and analyzed. The above mentioned financial indicators have been gathered for the 38 companies within the sample. Several steps have been taken to analyze the data. To start of linear regressions have been executed. As already briefly discussed Power is theorized as the independent variable whereas each different financial indicator as the dependent. In other words, it is investigated if Power is able to influence financial performance. Within this first step it is chosen to add each company six times to the database (6 years) with the financial ratio for each specific year but with the same Power score. This is done to extend the amount of observations and make the linear regression more robust. The tests for the other financial indicators are performed the same way and thus ROA can be substituted for the other ratios, see formula:
ROA
i, t= ∝ + 𝛽
1Power
i+ ∈
One can notice that in the correlations table provided in the Appendix Figure 3, the variables Age and the Number of Employees are noted. The theoretical framework did not provide handles to interpret these variables and that’s why they are not brought under attention within the results section.
After having identified possible directions it is chosen to divide the sample in two groups with companies having an above average score on Power and a below average score on Power. This is done to investigate if the place on the index has influence on the financial performance.
These two groups have been tested with non-parametric Mann-Whitney U tests and has enabled the author to identify possible differences between the two groups. The Grouping variable is again Power and the testing variable is respectively the financial performance ratio. This method puts the scores of the Power construct more in perspective. When lowering the amount of observations to N=38 which represents the real sample a Mann-Whitney U test is used to deepen the possible relationships.
After having assessed this, another Mann-Whitney U test is performed between the companies with the highest levels of Power and the five companies with the lowest levels of Power. After these tests distinctions have been made between the four different sectors. The next section will step by step explain the results in greater detail.
30
5.0 Results
To start off, table 1 shows the betas for each respective financial ratio with respect to Power. The hypothesis was firstly tested with means of multiple linear regression. H0 predicted that the degree of Power has no relationship with financial performance. The betas are all positive (small effect) but it was not possible to show a significant effect (table 1 & 2)3.
Table 1: beta coefficient of independent variable Power in relation to the financial ratio's.
ROE ROA ROC Solvency Turnover Growth Net Profit Margin Beta coefficient 0,071 0,114 0,081 0,091 0,052 0,074
Table 2: p values of the linear regression (Power as independent variable, financial performance indicators as the dependent)
ROE ROA ROC Solvency Turnover Growth
Net Profit Margin p-value 0,346 0,135 0,282 0,161 0,535 0,339
31
Now that the sign of Power has been discovered, the amount of observations is lowered to the real sample (N=38). This is done to strengthen the findings from the regression and assess the relationship between the level of Power and the financial performance of the family businesses in greater depth. Table 3 shows the scores on Power for the 38 businesses in the sample. The average scores for Power was 0,65. The sectors 1-4 stand respectively for the Food & Drinks, Construction, Industry and Consumer Goods. The scores of Power within the sample range from 0,26-1,00. The N in table 3 represents a particular company.
Table 3: Scores of Power for the entire sample N=38 from lowest to highest
N SECTOR POWER N SECTOR POWER
35 2 0,26 5 3 0,57 28 2 0,38 36 3 0,58 3 3 0,42 8 3 0,6 10 2 0,42 6 3 0,62 23 3 0,44 1 4 0,67 30 3 0,47 20 3 0,67 17 3 0,48 24 4 0,67 34 3 0,49 14 1 0,71 9 1 0,5 2 3 0,75 31 3 0,5 13 3 0,75 38 1 0,51 32 2 0,75 18 1 0,52 33 2 0,78 12 4 0,55 4 1 0,87 25 1 0,55 16 3 0,88 7 3 0,56 15 2 1 11 4 0,56 19 2 1 21 2 0,56 22 2 1 29 3 0,56 26 3 1 27 4 1 37 4 1
To evaluate if a higher degree of Power results in better financial performance two groups are created within the sample and are compared by a non-parametric Mann Whitney U-test as
32
discussed in the methodology section. Companies with a score below average are categorized as ‘low’ and companies with a score above average are categorized as ‘high’, see table 4 and 54. Table 4: Mean ranks of the financial performance indicators of businesses with low and high levels of Power
N MEAN RANK SOLVABILITY Low 22 19,02 High 16 20,16 Total 38 TURNOVER GROWTH Low 20 17,43 High 13 16,35 Total 33 NET PROFIT MARGIN Low 19 13,71 High 9 16,17 Total 28 ROA Low 20 16,30 High 13 18,08 Total 33 ROE Low 20 16,83 High 13 17,27 Total 33 ROC Low 20 16,23 High 13 18,19 Total 33
Table 4 shows that overall, companies with a high degree of Power score higher on all the financial indicators except for Turnover Growth. These findings are in line with the linear regression except that in the linear regression calculation Turnover Growth also shows a positive sign.
Table 5: p values show that there is no significant difference between the financial performance indicators for companies with high and low levels of Power
SOLVABILITY TURNOVER GROWTH
NET PROFIT MARGIN
ROA ROE ROC
SIG. (2-TAILED) 0,756 0,754 0,460 0,606 0,897 0,568
33
Although it looks like that high-Power results in higher financial performance it has not been possible to find significance at the 95% confidence interval level, see table 5.
The last test to assess if it is possible to reveal a positive significant relationship between Power and the financial performance indicators a distinction is made between the five businesses with the highest and lowest Power levels. This makes the distinction more extreme and these results can be found in table 6 & 7.
Table 6: Mean ranks of the financial performance indicators of the five businesses with highest and lowest Power levels
N MEAN RANK SOLVABILITY Low 5 6,4 High 5 4,6 Total 10 TURNOVER GROWTH Low 5 4,4 High 4 5,75 Total 9 NET PROFIT MARGIN Low 5 4,0 High 2 4,0 Total 7 ROA Low 5 4,2 High 4 6,0 Total 9 ROE Low 5 4,40 High 4 5,75 Total 9 ROC Low 5 4,2 High 4 6,0 Total 9
The results again show that the five businesses with the highest levels of Power have overall better financial performance then the companies with the lowest levels of Power but this test has not been
34
able to show a significant effect either. Solvency is lower for the five companies with the highest score on Power5.
Table 7: p values show no significant difference between the financial performance indicators for the five companies that score the highest and lowest on Power
SOLVABILITY TURNOVER GROWTH
NET PROFIT MARGIN
ROA ROE ROC
SIG. (2-TAILED) 0,347 0,462 1 0,327 0,462 0,327
There is a positive sign found for the linear regression (although small betas) and the Mann-Whitney U tests strengthen these findings by showing the higher overall Mean ranks. So, concluding both test types show contingent results and hypothesis H0 can be rejected.
H0: The degree of Power has no influence on firm performance within Dutch Businesses.
H1: The degree of Power has a positive influence on firm performance within Dutch Businesses.
35
5.1 Comparing sectors
The average scores for Power of the respective sectors can be found in table 8. The average score for Food & Drinks (1) is 0,61. Construction (2) 0,68. Industry (3) 0,61 and Consumer Goods (4) 0,74. The combination of share ownership and the ratios between family and non-family members in the management and governance boards (Power) is thus the highest within the Consumer Goods sector. To evaluate if the differences in Power result in better or worse financial performance comparison analysis is done.
Table 8: Averages of Power per sector
N Sector Power Average N Sector Power Average
Company 4 1 0.87 Company 16 3 0.88 Company 9 1 0.5 Company 17 3 0.48 Company 14 1 0.71 Company 20 3 0.67 Company 18 1 0.52 Company 23 3 0.44 Company 25 1 0.55 Company 26 3 1 Company 38 1 0.51 0.61 Company 29 3 0.56 Company 10 2 0.42 Company 30 3 0.47 Company 15 2 1 Company 31 3 0.5 Company 19 2 1 Company 34 3 0.49 Company 21 2 0.56 Company 36 3 0.58 0.61 Company 22 2 1 Company 1 4 0.67 Company 28 2 0.38 Company 11 4 0.56 Company 32 2 0.75 Company 12 4 0.55 Company 33 2 0.78 Company 24 4 0.67 Company 35 2 0.26 0.68 Company 27 4 1 Company 2 3 0.75 Company 37 4 1 0.74 Company 3 3 0.42
36 Company 5 3 0.57 Company 6 3 0.62 Company 7 3 0.56 Company 8 3 0.6 Company 13 3 0.75
To start of Food & Drinks and Construction is compared. These two sectors have the same average on Power. Table 9 shows the mean ranks of the Mann-Whitney U test and table 10 shows significance.
Table 9: Mean ranks of financial indicators and Power for Food & Drinks and Construction
N MEAN
RANK
POWER Food &
Drinks
6 7,17 Construction 9 8,56
Total 15
SOLVABILITY Food & Drinks 6 7,50 Construction 9 8,33 Total 15 TURNOVER GROWTH Food & Drinks 5 5,80 Construction 9 8,44 Total 15 NET PROFIT MARGIN Food & Drinks 5 7,60 Construction 7 5,71 Total 12 ROA AVERAGE Food & Drinks 5 9,40 Construction 9 6,44 Total 14 ROE AVERAGE Food & Drinks 5 10
37 Construction 9 6,11 Total 14 ROC AVERAGE Food & Drinks 5 9,60 Construction 9 6,33 Total 14
There is no significant difference found between the two sectors, at the 90% confidence interval level ROE was significantly higher within Food & Drinks. Solvability and Turnover Growth were higher within the Construction sector but found insignificant. Net profit Margin, ROA and ROC were higher within the Food & Drinks industry but no significance could be discovered. The difference in Power was not found significant.
Table 10: p values show that there is no significant difference between the financial performance indicators between the Food & drinks and Construction sector
Solv. Turnover
Growth
Net Profit
Margin ROA ROE ROC POWER
Sig.
(2-tailed) 0,724 0,257 0,372 0,205 0,096 0,162 0,554
The other industries are compared in a similar way but for the overall readability of the paper full tables are provided within Appendix Figure 12.
Food & Drinks and Industry did provide no significant difference in respect to the Power construct. Solvability, Turnover Growth and Net Profit Margin where higher for the Industry sector but insignificant. ROC was significantly higher for the Food & Drinks sector on the 90% confidence level. ROA and ROE where higher as well for Food & Drinks but was not able to show a significant difference.
When comparing Food & Drinks and Consumer Goods no significant differences is found. But it is interesting to see that all the Mean Ranks are higher for the Consumer Goods sector, with the exception of Solvability. The difference in Power is almost significant at the 90% level with a p- value of 0,126. Consumer Goods has the highest score on Power of all the industries and as just explained has for every financial performance indicator a higher Mean Rank (exception of Solvability).
38
Construction and the Industry sector did now show significant differences for Power and the financial performance indicators. The Mean Ranks for Power where in favor of Construction but all the financial performance indicators where in favor of the industry Sector.
When comparing the Construction and the Consumer Goods Sector, Power was not significantly different but Power had a higher Mean Rank within the Consumer Goods sector. ROE & ROC where significantly higher at the 95% confidence interval level within the latter sector which also scored (insignificantly) higher on Power.
When comparing Consumer Goods in respect to the Industry sector most results are in favor of Consumer Goods. Power was higher within Consumer Goods (not significantly) and Solvability was in favor for the Industry sector. ROA ROE & ROC showed a significant effect on the 95% confidence level in a favor of Consumer Goods6.
6 See Appendix Figure 12 for the test-statistics and Mann-Whitney U test output for all the different sectors
39
6.0 Conclusion & Discussion
The author endeavored to utilize the F-PEC index for the first time within Dutch businesses and made an effort to examine its relationship with financial performance. Initially, the respective index was operationalized due to the definitional disparity across studies that has been demonstrated. Therefore, it prompted the need for a revolutionary index that has the potential to elevate the confusing state of the family business (Rutherford & Kuratko, 2008). By utilizing the F-PEC and relating it to six different financial indicators, the author sought to answer the research question: Does the degree of Power have a positive influence on financial performance? The results show that it has not been possible to discover strong relationships for the Power construct in relation to the financial performance indicators. But it needs to be recognized that the Mann-Whitney U tests did provide insights, that within the sample used, the Mean Ranks where higher for companies with an above average (high) level of Power resulting in rejecting the H0 hypothesis
Still, one can speculate the positive influence of Power to be attributed to an influence of stewardship residing in the companies. Agency costs can therefore be considered to be low where opportunistic behaviour is not prominently present. Because of the positive (though insignificant) relationship, one could still argue that the lack of professionalism is not necessarily true, wherewith the same accounts for nepotism and managerial entrenchment.
40
The next section elaborates on findings regarding the whole F-PEC index. The overall goal of the latter questions within the questionnaire (see appendix) was to operationalize the Experience, Culture and familiness construct. As discussed Experience and Culture have been the topics of other researches using the same database.
Familiness is calculated as the weighted average of Power, Experience and Culture. Each business from the sample could score a familiness level from 0 to 1. Thus Power, Experience and Culture each account for ⅓ of the total familiness level.
If the scores found for Power (0,80), Experience (0,60) and Culture (0,40) are as follows, the total familiness level for the respective company would be: (1/3*0,80) + (1/3*0,60) + (1/3*0,80) = 0,726.
The results for the whole F-PEC index show stronger outcomes. To answer the question whether the F-PEC index suffices as the proverbial machete to cut through the theory jungle, the answer is: for the Dutch family businesses in the sample, not really.
To assess if there is a relation between the amount of familiness and financial performance a similar linear regression is performed as for the Power construct. The betas are shown in table 11
Table 11: Beta coefficients of the combination of Power, Experience and Culture which makes familiness
ROE ROA ROC Solvency Turnover Growth Net Profit Margin Beta coefficient 0,031 0,170 0,044 0,200 -0,012 0,212
As can be seen all the signs are positive (but small) with the exception of Turnover Growth. Table 12 shows that there is a significant positive relationship found for Solvency and Net Profit Margin but just as with the Power calculations, no distinction is made between high and low levels of familiness7. That’s why after the linear regression a similar Mann Whitney U-test is performed by
41
dividing the family businesses in two respective groups. This is done to strengthen the linear regression results.
Table 12: p values of the linear regression (familiness as independent variable, financial performance indicators as the dependent)
ROE ROA ROC Solvency Turnover Growth
Net Profit Margin p-value 0,743 0,068 0,639 0,025* (0,906) 0,013*
Family businesses with a below average score on familiness are coded as ‘low’ and those with an above average score as ‘high’. In table 13 the scores can be found for the familiness construct per family business. The range is from 0,4 till 0,71. The results in table 14show the mean ranks of the financial performance indicators for the respective groups.
Table 13: The scores for the familiness construct of the family businesses within the sample, plus the average
N SECTOR FAMILINESS N SECTOR FAMILINESS
18 1 0,4 6 3 0,72 35 2 0,5 14 1 0,72 2 3 0,53 32 2 0,73 15 2 0,57 34 3 0,73 23 3 0,57 16 3 0,74 21 2 0,59 28 2 0,74 11 4 0,61 3 3 0,76 9 1 0,64 1 4 0,77 12 4 0,64 7 3 0,77 20 3 0,66 29 3 0,77 30 3 0,66 13 3 0,78 36 3 0,66 37 4 0,78 31 3 0,67 24 4 0,8 38 1 0,67 26 3 0,81 17 3 0,68 33 2 0,81 8 3 0,69 4 1 0,86 10 2 0,69 22 2 0,86
42
25 1 0,71 19 2 0,98
5 3 0,72 27 4 0,99
Average: 0,71
Table 14: Mean ranks of the financial performance indicators of businesses with low and high levels of familiness
N MEAN RANK SOLVABILITY Low 18 17,64 High 20 21,18 Total 38 TURNOVER GROWTH Low 15 16,17 High 18 17,69 Total 33 NET PROFIT MARGIN Low 13 10,73 High 15 17,77 Total 28 ROA Low 15 14,70 High 18 18,92 Total 33 ROE Low 15 15,17 High 18 18,53 Total 33 ROC Low 15 14,63 High 18 18,97 Total 33
The results for familiness show stronger results then the calculations for the Power construct whereas every financial performance indicator has a higher mean rank for businesses with an above
43
average score on familiness. The differentiation between the groups resulted that Solvability does not show a significant relationship anymore8.
Table 15: Shows p values; a significant difference for Net Profit Margin between companies with high and low levels of familiness has been identified
SOLVABILITY TURNOVER GROWTH
NET PROFIT MARGIN
ROA ROE ROC
SIG. (2-TAILED) 0,327 0,651 0,024* 0,212 0,320 0,199
Finally, familiness shows a positive significant relationship with Net Profit Margin and indicates that companies with high levels of familiness perform significantly better then family businesses with lower levels of familiness for this financial indicator. The unique bundles and resources of the family firm could be the explanation for these higher Net Profit Margins. Concluding the separate construct of Power was not able to show convincing results, but the combination of Power, Experience & Culture enucleates more interesting insights. The findings for familiness in relation to Net Profit Margin could also be interpreted with means of shared family values, reciprocal altruism, shared family language, loyalty, trust and strong relationships translating in efficiency and effectiveness.
8 See Appendix Figure 19 for the descriptive statistics & test-statistics for the distinction between below and above
