Amsterdam Business School
A comparison between family firms and non-family firms regarding
financial performance measures and corporate sustainability
performance
Name: William Karsten Student number: 10193677 Thesis supervisor: A. Sikalidis Date: 24 June 2017
Word count: 12.388
MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam
2
Statement of Originality
This document is written by student William Karsten who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
The purpose of this research was to investigate whether family firms with a higher performance in corporate sustainability also have a higher financial performance.
The data sample was taken from the ASSET4 ESG Database from Thomson Reuters and contains a total amount of 874, 1483 and 1002 companies respectively in 2014, 2015 and 2016 which also respectively included 33, 37 and 26 family firms during those years. Determining whether a company was categorized as a family firm was done by using the GFBI, this list contained 101 biggest family firms in the US in terms of revenue and used a 30% threshold in ownership. The analysis contained univariate testing and multiple regressions including the tests for multiple regression assumptions.
The results indicated that family firms score significantly higher in corporate sustainability performance compared to non-family firms (H1). Also did the findings of this paper support that there is a positive relation between the corporate sustainability performance and the financial performances measures (H2) but found no relationship between family firms structure and these two performances when compared to non-family firms. (H2). Also was concluded that corporate sustainability performance becomes increasingly important during the period of analysis 2014 to 2016.
Limitations in this research are that mainly big corporations are being tested and only big family firms whom weren’t privately owned because it was not possible to find data on these companies while using the ASSET4 ESG Database. Another limitation is that only three years have been incorporated in this research. Third limitation for this research is that it is focused on financial performance and corporate sustainability performance and not the costs that can occur when investing in corporate sustainability performance.
Key words: corporate sustainability performance, financial performance, family firms and non-family firms.
4 Table of contents
1. Introduction 6
2. Theoretic background and hypothesis development 8 2.1 Corporate sustainability reporting 8 2.2 Corporate sustainable performance –
measurements and financial performance 10 2.3 Competitive advantage of family companies 12
2.4 Hypothesis development 13
3. Data sources and samples 14
3.1 Data sample sources 14
3.2 Data sample selection 15
4. Methodology and empirical model development 16 4.1 Proxy for determining family owned and –
non-family owned businesses 17
4.2 Proxies for measuring corporate –
sustainability performance 17
4.3 Proxies for measuring the performance – in corporate sustainability between family –
firms and non-family firms 18
4.4 Proxies for financial performance 18
4.5 Proxies for control variables 19
4.6 Empirical Model 19
5. Results 20
5.1 Univariate tests 21
5.2 Descriptive statistics for the regression 26 5.3 Testing linear regression assumptions 28
5.4 Multiple regression results 31
6. Conclusion and Discussion 35
7. References 39
5 List of tables and figures
Figure 1: Scatterplot residuals ROA 2014 Figure 2: Scatterplot residuals ROA 2015 Figure 3: Scatterplot residuals ROA 2016 Table 1: Group statistics univariate test 2014 Table 2: Univariate test 2014
Table 3: Group statistics univariate test 2015 Table 4: Univariate test 2015
Table 5: Group statistics univariate test 2016 Table 6: Univariate test 2016
Table 7: Descriptive statistics of the weighted-average scores
Table 8: Descriptive statistics of the dependent variables (financial performance measures) Table 9: Descriptive statistics of the control variables
Table 10: Durbin-Watson d-values
Table 11: Pearson’s Correlation Matrix 2014 Table 12: Pearson’s Correlation Matrix 2014 Table 13: Pearson’s Correlation Matrix 2014
Table 14: Results of the regression for ROA in 2014, 2015 and 2016 Table 15: Results of the regression for ROE in 2014, 2015 and 2016 Table 16: Results of the regression for FCF in 2014, 2015 and 2016
6 1. Introduction
Sustainability has become more relevant in the last two decades because it’s becoming more feasible that climate is changing. Politicians, corporations and the public have become increasingly aware of this since the latest climate reports repeatedly show that there is significant evidence that the climate is changing. Corporations are aware of this and know that a bad performance on sustainability could damage the company’s reputation which can possibly influence future performance. For instance, the environmental scandals with Volkswagen in September 2015 and the BP disaster in April 2010 have led to an enormous environmental and social damage (Dwyer, 2009). Furthermore is this also affecting the performance of these companies because of legal claims and other costs regarding environmental and social issues. Therefore corporate sustainability has become more important in structuring and governing an organization.
Corporate sustainability performance has become an important subject on international debates and gained progress by the 1992 Framework Convention on Climate Change, followed by the 1997 Kyoto protocol but these measures were regarded as inadequately formed and implemented, the progress on sustainability was minor but opened up further negotiations (Rajamani, 2016). Eventually at the 2011 Durban Conference was agreed to negotiate on a new climate agreement which would lead to the 2015 Paris Agreement (Rajamani, 2016). The ambition of this agreement is to lower greenhouse gas emissions. But more importantly the 2015 Paris Agreement made many parties all over the world to do concessions, in terms of differentiation between developed and developing countries, in order come to an agreement. Prior the 2015 Paris Agreement the United Nations launched the Global Compact Initiative in order to analyze corporate sustainability performance and the OECD organized conventions to introduce principles and guidelines which elaborated on corporate sustainability performance (Aras and Crowther, 2009).
The second aspect of this paper concerns family firms, because it is argued that family firms often have longer investment horizons leading to greater investment efficiency (James 1999). Furthermore family firms are better at mitigating unfavorable short sighted investments decision from managers (Stein, 1989). Too much risk taken now can damage the company in the future and because the large shareholder structure of family firms are they more likely to avoid these risks (Demsetz and Lehn, 1985; Naldi et al., 2007). Moreover it seems advantageous to have a large shareholder(s) whom is also active in the managerial
7
board in order to mitigate managerial expropriation because these large shareholder(s) can influence and monitor the firm. Anderson and Reeb (2003) argue that family firms outperform non-family firms and that firm performance is higher when there is a family member serving as a CEO compared with a non-family member CEO. In general it seems that family firms are more focused on the long term then in this paper it is expected that family firms also put more effort in tackling sustainability issues compared with non-family firms. Thus the main question of this research topic is: Do family companies with a higher performance in corporate sustainability also have a higher financial performance?
There is substantial research about family companies and corporate sustainability, but the comparison between non-family firms and family firms in the context of sustainability and financial performance researched much yet. The main question is relevant because sustainability plays a major role in nowadays life and will continue to do so. If investors and customers do care about environment, then companies who do a lot about sustainability is it also expected that they should have higher share values than companies who aren’t caring about environmental issues. This difference can be influenced by sales, decrease in reputation or direct claims from environmental damage. This paper aims to research if family firms perform better on sustainability topics. This will be done by analyzing public data and trying to find a relation between financial performance and corporate sustainability performance of family firms.
This paper will first give an overview of existing literature on corporate sustainability performance and financial performance regarding family firms. The first and current chapter provides an introduction to the subject and presents the main research question. The second paragraph provides prior literature on corporate sustainability and on how corporate sustainability performance is influencing financial performance. Moreover in this paragraph is also stated what it is that defines family firms and presents the relation between investors and the management of companies which influences the financial performance. Furthermore does the second paragraph present the two separate hypotheses that are being used in this research aiming to answer the main research question. In the third paragraph of this paper is explained on how the quantitative research is being approached and also shows an overview of which data sources and the data samples that will be used to test the hypotheses. The fourth paragraph of this research presents the methodology and shows how the models of this research have been developed. The methodology explains on
8
how two sets of models have been developed in order to test the two hypotheses. The hypotheses are being approached with univariate testing and multiple regressions. The results of these tests are being presented and analyzed in the fifth paragraph. The sixth and final chapter the conclusion is being drawn and the findings provide ground for discussion.
This research has analyzed whether family companies with a higher performance in corporate sustainability also have a higher financial performance. Evidence was found to support that family firms score significantly higher in corporate sustainability performance compared to non-family firms (H1). Further evidence supported that there is a positive relation between the corporate sustainability performance and the financial performances measures (H2) but found no relationship between family firms’ structure and these two performances when compared to non-family firms. (H2). Another conclusion that can be drawn from the results is that corporate sustainability is becoming increasingly important for companies in general. Especially that this research found a great difference in finding between 2014 and 2015.
2. Theoretic background and hypothesis development
First will be researched what corporate sustainability actually is and what the connection is between sustainability and financial performance. Moreover will the development of corporate sustainability during the last few decades be discussed. Thereafter the literature part about family firms will be discussed and explained what they actually are and what makes them a special group of interest worth being researched. Third, because family firms are apparently different than non-family firms the relation between these will be analyzed and explained where a large part will be done by the agency theory.
2.1 Corporate sustainability reporting
Brundlandt (1987) has defined sustainability in a global framework where he states that sustainability must try to satisfy the human needs in a continuous way to proceed to the ultimate goal in order for humanity to be sustainable. Dyllick and Hockerts (2002, p. 131) have transported this idea to a corporate level and they define corporate sustainability as: ‘’meeting the needs of a firm’s direct and indirect stakeholders without compromising its ability to meet the needs of future stakeholders as well.’’ In years prior to this research in of Dyllick and Hockerts (2002) there was a tendency to overstate short term gains of firms by
9
an increased focus on quarterly results. The focus on short term profits is contrary towards the goals and core of sustainability which is long term orientated. They stated that in the spirit of sustainability the firm’s have to meet the present needs and the future needs of its stakeholders. Short-term gains tend to be higher valued than long-term gains due to the economic discount rate. This causes managers to focus on the economic capital of their decisions and disregard long-term costs such as social- or environmental costs. Therefore firms need to manage also natural capital and social capital to achieve long-term sustainability.
Thus there are three key elements namely economic-, natural-, and social capital aspects. Furthermore need the short- and long-term aspects of these key elements be considered in order to give a clearer view of sustainability in general. Hicks (1946) explained how economic capital is sustainable due the use of income consumption. People are sustainable when they can consume a certain amount during a week and still own the same amount as at the beginning of the week. In short a person is viewed sustainable when he’s not spending more than he gains in a particular timespan. Dyllick and Hockerts (2002) conclude that both financial and management accounting are only able to provide an approximation of a firm’s economic capital. Intangible capital becomes more important when the gap between the book value and the market value of a firm increases. Therefore contains economic capital besides financial capital also tangible- and intangible capital which are needed to be managed to acquire economic sustainability. Finally they state that companies are economically sustainable when at any time cash flows are sufficient and provide a guaranteed liquidity to ensure shareholders receiving a persistent positive return.
A sustainable use of natural capital is based on ecological sustainability whereby is realized that earths’ natural capital can be depleted (Lovins et all, 1999, p. 146). Natural capital can be divided in two forms; natural resources and ecosystem services. Natural resources can be either renewable such as livestock, wood, water or it can be non-renewable such as fossil fuel or scarce metals as lithium. Natural capital can also be ecosystem services which are for example climate stability or plant and animal reproduction. Thus companies are ecologically sustainable when they consume less of natural capital than nature can reproduce. They do not exhaust more carbon dioxide or other gasses that contribute to global warming than nature’s capability to absorb these emissions. Finally they
10
do not participate in activities that damage the natural eco-system services (Dyllick and Hockerts, 2002).
The last key element is social capital which can be divided into two subcategories; human capital and societal capital. Human capital is about the capabilities of employees of a company which contains their skills, loyalty and motivation. Societal capital means the infrastructural environment around the firm, such as educational level, infrastructure, entrepreneurialism and other public services (Dyllick and Hockerts, 2002). To be a socially sustainable company, firms have to support local and national services in order to grow economically and socially. It’s in the own interest of companies to invest public services like the educational and democratic system for instance because a stable social environment is important for a firms future growth and future in general (Dyllick and Hockerts, 2002). But generates costs for firms and can therefore lower profits which are unpleasant for shareholders. Therefore firms face a trade-off between stakeholders and need to find a balance which the stakeholders can agree with (Kaptein and Wempe, 2001). To conclude companies are socially sustainable when they add value to communities which they operate in by increasing the skills and public services of these communities as well as managing social capital so that stakeholders can agree with companies decisions (Dyllick and Hockerts, 2002). Whilst Dyllick and Hockerts (2002) only mention three main capitals does this paper address a fourth capital namely corporate governance. Corporate governance is a set of structures that provide boundaries for the organizations which affect employees, managers, investors and other stakeholders (Starks, 2009). Corporate governance is of importance because the structure of companies and decision making process influences on how companies are performing. In this paper the does address the corporate governance capital because the influence of family firms on corporate sustainability performance and financial performance is being analyzed. Thus factors that influence the corporate governance of companies is the structure of the board of directors and on how they are being monitored by investors. (Starks, 2009; Chrisman et al., 2005).
2.2 Corporate sustainable performance measurements and financial performance
Recent literature has elaborated the three capitals and accordingly to Atriach et al. (2010) does corporate sustainability performance measure the impact on society by examining a firm’s environmental, social and economic factors. The focus of their study is the
11
relationship between economic factor, which the authors call the corporate financial performance, and the corporate sustainability performance. While using a stakeholder framework they investigate what factors influence a firm’s decision to invest in corporate sustainability performance. They’ve found that size, profitability and potential for growth are positively related to high investments of firms in corporate sustainability performance. Furthermore they conclude that some types of firms invest more in corporate sustainability performance to maintain a market position.
Other studies argue that investment in corporate sustainability performance is costly and negatively associated with a firm’s economic performance (Becchetti et al, 2005). Firms that invest much in corporate sustainability performance incur high costs with regard to environmental issues, employee conditions, donations, societal improvements and increasing the standard of living for underdeveloped communities. This approach implies that capital and resources are re-allocated to non-financial investments which lead to lower financial returns for the firm’s stakeholders.
Secondly another study suggests that there is no direct relationship between corporate sustainability performance and a firm’s corporate financial performance because the relationship is too complex and is influenced by many factors (Ullmann, 1985). According to him is this complex relation too difficult to comprehend and therefore is it unlikely to find significant evidence for supporting a direct relationship between corporate sustainability performance and financial performance.
Thirdly some studies claim a positive association between corporate sustainability performance and financial performance. Barnett (2007) found, in contrary to Becchetti et al. (2005) that investments in corporate sustainability performance outperform the incurred investments costs which agrees with the study of Artiach et al (2010). They argue that it is beneficial for firms to invest in employee conditions, improved relationship with stakeholders, goodwill and increased access for gaining capital. Furthermore does the resource view give another insight into the relationship between corporate sustainability performance and financial performance (Clarkson et al, 2011). The resource view argues that firms whom have many resources to spare they’ll be more willing to invest into corporate sustainability performance simply because they can afford it. Thus only firms that are bigger and have more resources to invest in corporate sustainability performance will therefore
12
have a higher underlying financial performance to afford the investments in corporate sustainability performance.
2.3 Competitive advantage of family companies
At present day is the major part of most developed economies being dominated by family firms. Roughly one-third of the S&P 500 consists of family firms which are active in a wide range of industry sectors (Astrakhan and Shanker, 2003). The fact that the majority of the firms are family firms, has led to an increased research effort into family firms which its peak halfway the first decade of this century when major publications were issued on this subject. Prior to this period researchers argued about what amount of family ownership defines a family firm. A theoretic framework has been developed to integrate different approaches like the strategic management approach and the agency theory.
Several studies have concluded that family companies have certain competitive advantages against non-family owned companies. Family firm have been described as a friendly working environment which increases the employee’s loyalty and care for the firm (Ward 1988). Moreover Habbershon and Williams (1999) argue that family firms pay better wages and employees perform better because of the intrinsic value of their familiarity with the firm’s owners. Also they offer more flexible working places, lower recruitment costs and are more effective in labor intensive complex projects.
Aronoff and Ward (1995) have found that family firms have a better reputation in terms of trustworthiness and also incur lower transaction costs. This is supported by Minichilli and Corbetta (2010) they argue that firms have higher financial performance when there are many family members on the management board because they have lower agency costs. Another important characteristic is that family firms have lower agency costs, because of the overlapping principle-agent relationships within the family. It is argued that this relationship gives family firms better developed goals on the long-term vision and a stronger sense of mission compared with non-family firms (Miller and Breton-Miller, 2006; Schulze et al., 2001). While Tagiuri and Davis (1996) have found indication that family firms have better decision-making processes, less organizational structure and reduced monitoring costs.
Also family firms tend to reflect the personal beliefs and values of the founder/family. Family firms show greater commitment to the founder’s beliefs and put more emphasize in their missions to these values (Moscetello, 1990). He elaborates that
13
family firms have more room for self-criticism and adapt easier to fast changing conditions by pressing major changes without harming production. Also do family firms pay more attention to research and development and seem to have a more creative internal environment compared with non-family firm (Naldi et al, 2007).
Most important is that family firms seem to perform better when analyzing financial performance measured when they are compared with non-family firms (Astrachan et al., 2003). Some studies found that the return of assets, return of equity and other financial performance measures are positively related with being a family firm while family firms also are taking less risk (Naldi et al, 2007; Zahra et al., 2004; Anderson and Reeb, 2003; Minichilli and Corbetta, 2010)
2.4 Hypothesis development
Based on the literature mentioned in the literature review section there are hypotheses formed. These hypotheses will be further extended to give extended knowledge of the subject containing family firms and their performance on corporate sustainability. The core of this paper is researching if family firms score significantly higher in corporate sustainability performance when being compared with non-family firms. Previous research explains that corporate sustainability performance is determined by four main determinants namely; economic performance, corporate governmental performance, social performance and environmental performance. This paper analyzes whether family firms perform better on sustainability performance compared to non-family firms, therefore a fourth determinant will be firm type. Because family firms seem to be more long term viewed is it expected they will also perform better on corporate sustainability. Corporate sustainability seems to become more important since recent developments like the 2015 Paris Agreement and therefore it is also expected that corporate sustainability performance will increase over time. In order to analyze which type of firm scores better on corporate sustainability performance formulates the first hypothesis:
H1: Family firms score significantly higher in corporate sustainability performance compared to non-family firms.
14
Secondly this paper will go more in depth into how the ownership structure influences a firm’s financial performance. Therefore is this research analyzing the relationship between corporate sustainability performance and their financial performances. This is executed by comparing the largest family firms in terms of revenue in the US to the largest non-family firms on how they perform on corporate sustainability and financial measures. It is expected that firms whom perform better on corporate sustainability also perform better on financial performance measures and that being a family firms shows a positive relationship on these financial performance measures. Therefore the second hypothesis is:
H2: There is a positive relation between the corporate sustainability performance and the financial performances measures of family firms compared to non-family firms.
These two hypotheses form the backbone on which the hypotheses will be elaborated with proxies to test data gathered from the ASSET4 ESG database and eventually to develop the models for the tests. Because there has been quite some development on corporate sustainability performance in recent years, especially the Paris Climate Agreement, has made corporate sustainability performance development over time an important part for this research. Therefore the setup of the research will provide data for years 2014, 2015 and 2016. A more thorough description of the research setup will be provided in the following section, data and method, of this paper.
3. Data and method
In order to test the hypothesis will this research contain a quantitative database research whereby family firms will be compared with non-family firms on how they perform on corporate sustainability. Data form the ASSET4 ESG database provided Thomson Reuters and the Global Family Business Index (GFBI) is selected and is focused on the US market during a period from 2014 to 2016.
3.1 Data sample sources
This research will be a quantitative database research where will be first determined which firms are considered family firms. The Global Family Business Index has compelled the top
15
500 privately and publicly listed companies worldwide in terms of total revenue and has determined whether they are family firms or not. They’ve used a 50 percent ownership threshold for unlisted companies and a 30 percent ownership threshold for companies with a listing on a stock exchange. This list has been developed by the renowned University of St. Gallen in Switzerland in cooperation with EY’s Global Family Business Center of Excellence. This research is limited to the US market and therefore the amount of family firms is 101 family firms. These companies have been ranked by their total revenues and contain both publicly as privately owned firms.
Furthermore the ASSET4 ESG database has compelled a series of listed companies worldwide and how they perform on corporate sustainability, but in order to limit the scope of this research will only the data from US listed firms be used. This database has used social, environmental, corporate governance and economic performance as pillars and will be used to measure how companies perform on corporate sustainability performance. This research will use the firms that have a listing in both indexes in order to analyze the relative performance of family firms against non-family firms.
Using the ASSET4 ESG database and GFBI index will eliminate personally biased ratings because the determinants of this the GFBI index is predetermined to exclude personal bias and hence will give a clearer view than if the firms will be individually ranked on how they perform. Corporate sustainability performance is analyzed by using the economic, social and environmental efforts of firms as determinants and also industry specific comparison will be made between the non-family firms and family firms.
3.2 Data sample selection
In 2017 the data sample for US listed firms in the ASSET4 ESG database contains 1755 firms. From this database the four main pillars; social, environmental, corporate governance and economic performance for the years 2014, 2015 and 2016 are used to determine the corporate sustainability performance of firms. In order to provide a comprehensive view on corporate sustainability performance and limit variables in the regression have these four pillars formed a weighted average score on corporate sustainability performance for these companies. For these same companies the financial data; return on assets, return on equity, financial cash flows, asset growth and revenues have been collected for the years 2014, 2015 and 2016.
16
Every year the ASSET4 ESG database increases in the amount of firms they are assessing and therefore there was no corporate sustainability performance data for 824 companies, from the total of 1755 companies, in 2014 and for an additional 57 companies there was no financial data. The amount of companies where there was no corporate sustainability performance data found in 2015 was 154 and additionally for 118 companies there was no financial data. 2016 had no corporate sustainability performance data for 662 companies and for an additional 91 companies there was no financial data. Remarkably 2016 contained less data compared with 2015 on both corporate sustainability performance scores as on financial performances. This is likely that Thomson Reuters was still assessing on how these firms were performing on corporate sustainability performance or financial performance or there was no data yet released by these companies.
Moreover from the 101 family firms in the US found in the GFBI index there was no corporate sustainability performance and/or financial data found for some firms because they were privately owned or for other unknown reasons. Therefor the amount of family firms included in the research for the period between 2014 and 2016 contained 33 family firms in 2014, thus for 68 family firms there was no corporate sustainability performance or financial data found. In 2015 the amount of researched family firms after filtering was 37 and in 2016 this was 26. Remarkably for 2016 this again lower than 2015 but the same reasons mentioned before are assumed to be the origin of this.
Also this research does also not include control variables for the different years. An important part of this research is to give an insight in how performance for corporate sustainability develops over time.
4 Methodology and model
In this chapter the methodology and models of this research is explained. First are the proxies presented, these proxies have been taken from existing literature and will form the basis for the development of the models later in this chapter. The first paragraph in this chapter will provide on how categorization of family firms is being determined. The second paragraph provides proxies for measuring performance on corporate sustainability. The third paragraph of this chapter presents the proxies for measuring the relationship of corporate sustainability performance between family firms and non-family firms. The fourth paragraph presents the proxies for financial performance. Furthermore the fifth paragraph presents the
17
proxies for control variables. At last does the sixth paragraph present the empirical models which have been derived from the previous five proxies mentioned earlier.
4.1 Proxy for determining family owned and non-family owned businesses
Because this paper is assessing if family firm outperform non-family firms must first be determined whether a firm is a family firm or not. The proxy for determining if firms are family owned is taken from the GFBI. The GFBI uses a 30 percent threshold of family owned ownership and for publicly and privately owned firms from the GFBI. Though the data sample does not contain any privately owned firms, this is caused by the fact that the ASSET4 ESG Database did not contained data on these firms. Therefore is the amount of family firms reduced from 101 to 33, 36 and 26 in respectively 2014, 2015 and 2016.
(1) P>0.3
‘P’ represents the 30% threshold used by the GFBI to determining whether a firm is a family firm or a non-family firm. Later models in this chapter will develop this proxy into CSPF and DFAMILY.
4.2 Proxies for measuring corporate sustainability performance
In order to test the first hypothesis must be determined how firms score on corporate sustainability performance which has been split in four different proxies in order to calculate the scores of the two populations. These are Corporate Social Performance (CSOP), Corporate Environmental Performance (CENP), Corporate Governance Performance (CGOP) and Corporate Economic Performance (CECP) (Dyllick and Hockerts, 2002). The scores on these four factors are gathered in the ASSET4 ESG Database which provides for the data and provides the proxy:
(2) CSPi;t = α1 + β1 CSOPi;t + β2 CENPi;t + β3 CGOP;t + β4 CECPi;t + ε
CSPi;t is the score in corporate sustainability performance where i is the CSP score for firm i
in year t. Eventually will the CSP be developed into a dummy variable in the model for which it can take the value 1 for being a family firm according the GFBI and ASSET4 ESG Database
18
and 0 otherwise. The values from the following proxies will be taken from the ASSET4 ESG Database: CECPi;t is the Corporate Economic Performance for firm i in year t. CSOPi;t is the
Corporate Social Performance for firm i in year t, CENPi;t is the Corporate Environmental
Performance for firm i in year t and CGOPi;t is the Corporate Governance Performance for
firm i in year t.
4.3 Proxies for measuring the performance in corporate sustainability between family firms and non-family firms.
By further elaborating the proxy from Dyllick and Hockerts (2002) and combining this with the first proxy, which determines whether a firm is a family firm or not, provides the following proxy for measuring the performance in CSP score between family firms and non-family firms. There has been determined what a non-family firm is and which are non-non-family firms (1) according to the GFBI. CSP (2) scores from the ASSET4 ESG database are used to determine theCSOP, CENP, CCOP and CECP scores which generates the third and final proxy in order to test the first hypothesis.
(3) CSPFi;t = α1 + β1 CSOPi;t + β2 CENPi;t + β3 CCOPi;t + β4 CECPi;t + β5 CSP_WAi;t + ε
Where CSPFi;t is the score in performance of family firms and is measured as a 0/1 dummy
variable where 1 if firm it represents a family firm from the GFBI or 0 if representing a non-family firm. CECPi;t is the Corporate Economic Performance for firm i in year t. CSOPi;t is the
Corporate Social Performance for firm i in year t and CENPi;t is the Corporate Environmental
Performance for firm i in year t. CSP_WAi;t is the weighted average from the ASSET4 ESG
database and is comprised of the other four variables mentioned in this proxy
4.4 Proxies for financial performance
In order to address the second hypothesis it is necessary to calculate financial performance of family firms compared with the score of non-family firms which is derived from the model of Atriach et al. (2010) and combined with proxy (4)
19
Where FP;t is the score in Financial Performance with i is the FP score for firm i in year t.
ROAi;t is the return on assets measured as EBIT divided by total assets for firm i in year t.
ROEi;t is return on equity measured as EBIT divided by total equity for firm i in year t. FCFi;t is
financial capacity measured as free cash flow divided by net sales for firm i in year t. AS_GROWTH is ratio in asset growth for firm i in year t.
4.5 Proxies for control variables
Control variables used in this model is size and is measured in total revenues, furthermore three years are included in the data sample and are used to test the second hypothesis but is not included as a control variable but every year is tested separately.
4.6 Empirical Model
The proxies (1) and (3) mentioned above are used to develop and determine the first set of models for univariate testing in this research and address the first hypothesis. In order to test if family firm perform better on corporate sustainability the following models for the data sampling period of 2014-2016 is provided:
CSPF2014 = α1 + β1 CSOP2014 + β2 CENP2014 + β3 CCOP2014 + β4 CECP2014 + β5 CSP_WA2014 + ε
CSPF2015 = α1 + β1 CSOP2015 + β2 CENP2015 + β3 CCOP2015 + β4 CECP2015 + β5 CSP_WA2015 + ε
CSPF2016 = α1 + β1 CSOP2016 + β2 CENP2016 + β3 CCOP2016 + β4 CECP2016 + β5 CSP_WA2016 + ε
Where
CSPF= Dummy variable for determining family firms; determined from the GFBI CSOP = Corporate social performance; provided by the ASSET4 ESG Database
CENP= Corporate environmental performance; provided by the ASSET4 ESG Database CCOP = Corporate governance performance; provided by the ASSET4 ESG Database CECP = Corporate economic performance; provided by the ASSET4 ESG Database
CSP_WA = Corporate Sustainable Performance; the weighted average form the social, environmental, governance and economic pillars; provided by the ASSET4 ESG Database
In the following and final set of models which are derived from the proxy (4) combined with proxy (3) and used to perform multiple regressions. This set of models is focussed on
20
addressing the second hypothesis but also contains findings that complement to the first set of models which are addressing the first hypothesis. In order to test whether family firms outperform on corporate sustainability and financial performance compared with nonfamily firms, includes the control variables and uses the sampling period of 2014-2016 is provided:
FP2014 = α1 + β1 CSP_WA2014 β2 DFAMILY2014 + β3 SIZE2014 + ε
FP2015 = α1 + β1 CSP_WA2015 β2 DFAMILY2015 + β3 SIZE2015 + ε
FP2016 = α1 + β1 CSP_WA2016 β2 DFAMILY2014 + β3 SIZE2016 + ε
Where FPi:t is ROAi;t, ROEi;t, FCFi;t and AS_GROWTH
ROA = Return on Assets; net income / total assets ROE = Return on Equity; net income / total equity FCF = Free cash flows; net income / total sales AS_GROWTH = the ratio in asset growth in year t.
CSP_WA = Corporate Sustainable Performance; the weighted average form the social, environmental, governance and economic pillars
DFAMILY = Dummy variable coefficient determining for family firms SIZE = Revenues; control variables for the size of the firms
5 Results
In this chapter are the results, which have been obtained with the statistical data program SPSS, for the models being presented. The two hypotheses will be tested by the two sets of models provided in the methodology section. Firstly is the first set of models being tested with univariate tests for 2014, 2015 and 2016 in order to provide evidence for the first hypothesis. The results for second set of models and the second hypothesis will be presented by providing the descriptive statistics for the regression model in the second chapter. The third chapter will present the tests for the assumption which are needed to perform a linear regression for the second hypothesis. In the fourth and final chapter of this section provides the results of the regression from the second set of models which is used to provide evidence for the second hypothesis of this research.
21 5.1 Univariate tests
In this chapter the results for the unvariate tests which are applied to the first set of models; CSPFi;t = α1 + β1 CSOPi;t + β2 CENPi;t + β3 CCOPi;t + β4 CECPi;t + β5 CSP_WAi;t + ε. where
CSP_WAi;t is the weighted average of the four other pillars. This set of models is used to
obtain evidence for the first hypothesis that: family firms score significantly higher in corporate sustainability performance compared to non-family firms. The dependent variable is CSPF which is a dummy and is 1 when it is a family firm and 0 when otherwise.
Group Statistics CSPF N Mean Std. Deviation Std. Error Mean CSOP_SCORE 1 33 49,6988 31,09635 5,41318 0 841 47,7967 27,80203 ,95869 CENP_SCORE 1 33 45,9164 36,21292 6,30386 0 841 46,7065 31,98694 1,10300 CCOP_SCORE 1 33 71,7927 15,05294 2,62038 0 841 72,8547 17,85984 ,61586 CECP_SCORE 1 33 58,2524 24,39927 4,24737 0 841 54,9831 27,20261 ,93802 CSP_WA 1 33 56,1439 31,52743 5,48822 0 841 56,0620 28,24565 ,97399
Table 1: Group statistics univariate test 2014
Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper CSOP_SCORE Equal variances assumed 1,538 ,215 ,384 872 ,701 1,90211 4,95642 -7,82581 11,63002 Equal variances not assumed ,346 34,038 ,731 1,90211 5,49742 -9,26954 13,07375 CENP_SCORE Equal variances assumed 4,524 ,034 -,138 872 ,890 -,79016 5,70567 -11,98861 10,40828 Equal variances not assumed -,123 33,988 ,902 -,79016 6,39963 -13,79594 12,21561
22 CCOP_SCORE Equal variances assumed 2,053 ,152 -,337 872 ,736 -1,06201 3,15252 -7,24942 5,12541 Equal variances not assumed -,395 35,629 ,696 -1,06201 2,69178 -6,52316 4,39915 CECP_SCORE Equal variances assumed 1,161 ,282 ,680 872 ,497 3,26931 4,81003 -6,17128 12,70990 Equal variances not assumed ,752 35,194 ,457 3,26931 4,34971 -5,55934 12,09796 CSP_WA Equal variances assumed 3,722 ,054 ,016 872 ,987 ,08199 5,03504 -9,80022 9,96420 Equal variances not assumed ,015 34,046 ,988 ,08199 5,57398 -11,24513 11,40911
Table 2: Univariate test 2014
To determine if family firms score significantly higher in corporate sustainability performance compared to non-family firms in 2014 an univariate test is performed for the following model; CSPF2014 = α1 + β1 CSOP2014 + β2 CENP2014 + β3 CCOP2014 + β4 CECP2014 + β5
CSP_WA2014 + ε. Where CSPF is the grouping variable and is a 1 when it is a family firm and a
0 when otherwise. The amount of family firms in 2014 was 33 and the amount of non-family firms was 841. In 2014 table 2 presents that none of the variables seems to be significant with a significance level of 5 percent. Thus the univariate test for 2014 does not support evidence for a positive relation between being a family firm and their performance on corporate sustainability. Group Statistics CSPF N Mean Std. Deviation Std. Error Mean CSOP_SCORE 1 37 51,2130 31,06603 5,10722 0 1446 37,8805 29,99165 ,78871 CENP_SCORE 1 37 42,4922 34,15015 5,61425 0 1446 37,4209 30,47698 ,80147 CCOP_SCORE 1 37 67,4546 24,52052 4,03115 0 1446 61,7198 23,90587 ,62867 CECP_SCORE 1 37 61,9332 26,54198 4,36347 0 1446 40,7475 32,41133 ,85234
23
CSP_WA 1 37 55,8765 30,84708 5,07123
0 1446 42,8602 32,32033 ,84995
Table 3: Group statistics univariate test 2015
Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
F Sig. t df Sig. (2-taile d) Mean Differenc e Std. Error Differenc e 95% Confidence Interval of the Difference Lower Upper CSOP_SCORE Equal variance s assumed ,163 ,686 2,668 1481 ,008 13,33245 4,99770 3,52912 23,13579 Equal variance s not assumed 2,580 37,737 ,014 13,33245 5,16777 2,86846 23,79644 CENP_SCORE Equal variance s assumed 3,368 ,067 ,996 1481 ,319 5,07123 5,08982 -4,91279 15,05525 Equal variance s not assumed ,894 37,482 ,377 5,07123 5,67117 -6,41467 16,55713 CCOP_SCOR E Equal variance s assumed ,181 ,671 1,440 1481 ,150 5,73477 3,98258 -2,07733 13,54688 Equal variance s not assumed 1,406 37,772 ,168 5,73477 4,07988 -2,52614 13,99569 CECP_SCORE Equal variance s assumed 17,33 8 ,000 3,942 1481 ,000 21,18571 5,37448 10,64330 31,72812 Equal variance s not assumed 4,765 38,798 ,000 21,18571 4,44594 12,19145 30,17997 CSP_WA Equal variance s assumed 1,237 ,266 2,422 1481 ,016 13,01627 5,37515 2,47255 23,55999 Equal variance s not assumed 2,531 38,050 ,016 13,01627 5,14196 2,60737 23,42518
24
Table 4 presents the results for the model; CSPF2015 = α1 + β1 CSOP2015 + β2 CENP2015 + β3
CCOP2015 + β4 CECP2015 + β5 CSP_WA2015 + ε. This provides supporting evidence for the
hypothesis that family firms score significantly higher in corporate sustainability performance compared to family firms for 2015. The amount of family firms and non-family firms was highest in 2015 which counted 37 non-family firms and 1446 non non-family firms. The CSOP, CECP and WA variables are significant with a 5 percent significance interval and therefore suggest that there is a positive relationship with being a family firm and having a higher performance in corporate sustainability. Especially CECP appears to be very significant. This is promising for the second part of the results for this research. CECP contained the economic performance indicators of companies in the ASSET4 ESG Database and having a significant p-value is positive indication that the financial performance measures will be positive in 2015. More important is that the weighted-average score of the family firms is significant. The CSP_WA is a weighted-average of the four main pillars, thus in general the significance of CSP_WA suggests that there appears to be a positive relation between family firms and their corporate sustainability performance.
Group Statistics CSPF N Mean Std. Deviation Std. Error Mean CSOP_SCORE 1 26 43,8688 26,95321 5,28596 0 976 29,7458 24,97937 ,79957 CENP_SCORE 1 26 38,7538 33,07791 6,48711 0 976 28,0921 24,14017 ,77271 CCOP_SCORE 1 26 61,1185 24,84510 4,87253 0 976 57,8011 22,31166 ,71418 CECP_SCORE 1 26 56,5954 24,91738 4,88670 0 976 37,2118 30,28540 ,96941 CSP_WA 1 26 50,1800 26,74120 5,24438 0 976 35,1526 27,81205 ,89024
Table 5: Group statistics univariate test 2016
Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
F Sig. t df Sig. (2-taile Mean Differenc e Std. Error Differenc e 95% Confidence Interval of the Difference
25 d) Lower Upper CSOP_SCORE Equal variance s assumed ,385 ,535 2,839 1000 ,005 14,12308 4,97386 4,36267 23,88348 Equal variance s not assumed 2,642 26,157 ,014 14,12308 5,34609 3,13724 25,10892 CENP_SCORE Equal variance s assumed 11,69 9 ,001 2,199 1000 ,028 10,66173 4,84926 1,14584 20,17761 Equal variance s not assumed 1,632 25,714 ,115 10,66173 6,53297 -2,77425 24,09770 CCOP_SCOR E Equal variance s assumed ,755 ,385 ,746 1000 ,456 3,31741 4,44686 -5,40883 12,04365 Equal variance s not assumed ,674 26,085 ,506 3,31741 4,92459 -6,80362 13,43843 CECP_SCORE Equal variance s assumed 5,423 ,020 3,234 1000 ,001 19,38360 5,99369 7,62194 31,14526 Equal variance s not assumed 3,891 27,005 ,001 19,38360 4,98193 9,16163 29,60558 CSP_WA Equal variance s assumed 1,113 ,292 2,722 1000 ,007 15,02736 5,52135 4,19260 25,86211 Equal variance s not assumed 2,825 26,461 ,009 15,02736 5,31940 4,10243 25,95228
Table 6: Univariate test 2016
Table 6 presents the results for the model; CSPF2016 = α1 + β1 CSOP2016 + β2 CENP2016 + β3
CCOP2016 + β4 CECP2016 + β5 CSP_WA2016 + ε. In 2016 the amount of family firms counted 26
and the amount of non-family firms was 976. Similar as in 2015 the CSOP, CECP and WA variables are again significant with a 5 percent significance interval but also CENP is now significant, this suggests that the environmental pillar has become more important since 2015 which is in agreement with the expectations that corporate sustainability performance will increase over time. Also CECP is very significant again which makes the high significance level of 2015 in economic performance less likely to be a fluke. The weighted-average score
26
is again presenting evidence that is suggesting that there is a positive relationship with being a family firm and having a higher performance in corporate sustainability. In short 2015 and 2016, contrary to 2014, are providing evidence for the hypothesis that family firms score significantly higher in corporate sustainability performance compared with non-family firms. The only variable which hasn’t been significant is CCOP, which was the corporate governance pillar of the ASSET4 ESG Database. This suggests that the structure of governance in a company does not influence the performance on corporate sustainability which is very contrary to that being a family firms is a factor of corporate governance.
5.2 Descriptive statistics for the regression
In this chapter the descriptive statistics for the regression are being presented in order to provide an enhanced understanding of the second set of the models; FPi;t = α1 + β1 CSP_WAi;t
β2 DFAMILYi;t + β3 SIZEi;t + ε. With these models used in the regression the second
hypothesis is tested and tries to provide insight to the question whether there is a positive relation between the corporate sustainability performance and the financial performances measures of family firms compared to non-family firms.
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
CSP_WA 2014 874 3,29 96,97 56,0650 28,35654
CSP_WA 2015 1483 3,57 96,18 43,1850 32,33826
CSP_WA 2016 1002 5,26 96,39 35,5426 27,87457
Where for DFAMILY ''N'' contains 33 family firms in 2014, 37 in 2015 and 26 in 2016 Table 7: descriptive statistics of the weighted-average scores.
Table 7 presents the results for the CSP_WA which represents the weighted-averages scores on corporate sustainability performance. Observing the mean of the descriptive statistics is showing a decline from 2014 to 2016 and seems to have a relatively steady standard deviation. Given our previous findings in the 5.1 section of this paper that performed univariate testing on the CSP_WA rose the expectation that the mean of the CSP_WA would increase. A possible but unfounded explanation for this decline could be that in 2014 the companies that reported data in the ASSET4 ESG database were early adaptors and therefore were already more focused on corporate sustainability performance and that later on other companies joined which weren’t heavily focused on corporate sustainability
27
performance yet. The increase in minimum and maximum from 2014 to 2016 seems to indicate an increased effort in corporate sustainability performance from companies which agrees with the previous findings from this research. Table 7 also presents the amount of family firms in each year.
Descriptive Statistics Dependant Variables
N Minimum Maximum Mean
Std. Deviation ROA 2014 874 -140,05 47,45 5,7551 10,93423 ROA 2015 1483 -128,12 92,10 3,7151 12,67364 ROA 2016 1002 -68,81 99,05 4,6092 9,96014 ROE 2014 874 -7067,81 1325,34 7,0524 257,97482 ROE 2015 1483 -542,00 1479,81 12,4115 75,59086 ROE 2016 1002 -3933,07 1052,75 5,7619 149,04245 FCF 2014 874 -332,45 664,33 20,1148 36,38799 FCF 2015 1483 -206408,20 220,76 -202,4645 5685,97002 FCF 2016 1002 -331483,33 362,95 -339,5525 10496,04671 AS_GROWTH 2014 874 -77,08 431,83 10,3783 34,53486 AS_GROWTH 2015 1483 -90,96 7678,37 18,9997 207,87843 AS_GROWTH 2016 1002 -73,29 436,08 11,1796 33,58614
Table 8: descriptive statistics of the dependent variables (financial performance measures).
Table 8 gives an overview of the financial performance measures ROA, ROE, FCF and AS_GROWTH in the period between 2014 and 2016. The total amount of companies in 2014, 2015 and 2016 were respectively 874, 1483 and 1002 which also respectively included 33, 37 and 26 family firms. The financial variables are all ratios and show mostly positive means and standard deviations do not appear to differ much. Though the FCF variable appears to be volatile and is the only ratio which shows a negative for its mean and shows a big increase in its standard deviation.
Descriptive Statistics Control Variables
N Minimum Maximum Mean Std. Deviation
SIZE 2014 874 14689 485651000 13062401,77 30147316,850
SIZE 2015 1483 61 482130000 8300927,11 23390410,970
SIZE 2016 1002 78 485873000 5992314,68 22311202,545
Table 9: descriptive statistics of the control variables.
Table 9 shows the only control variable used in this regression, usually there will also be controlled for the yearly fluctuations with a year dummy, though in this setting every year is being evaluated separately to get a better insight in yearly developments which are
28
important since the increase of interest in corporate sustainability performance. Control variable SIZE is measured in total amount of revenue and there is a remarkable distinction between 2014 and the other two years. The minimum and mean are much higher in 2014 compared with 2015 and 2016, this is probably because of that the ASSET4 ESG database firstly incorporated the bigger companies and later on also included smaller US listed firms.
5.3 Testing linear regression assumptions
In the next chapter the assumptions which are needed for a regression are being explained tested. In order to use linear regression five assumptions must be satisfied; the type of variable, homoscedasticity, linear relationship, no auto-correlation and there must be no multi-correlation. The fallowing tests will relate to the second hypothesis and will focus the second model during 2014 till 2016; FP i;t = α1 + β1 CSP_WA i;t + β2 DFAMILY i;t + β3 SIZE i;t + ε
1. Type of variable
The dependent variables in the assumption need to be continuous or ordinal, this means that the variables need to be measured at a ratio or interval scale. The four different dependable variables in the second model for financial performance; ROAi;t, ROEi;t, FCFi;t and
AS_GROWTHi;t are continuous and thus the assumption is satisfied for the dependable
variables.
2. Homoscedasticity/Hetroscedasticity
For the second assumption of homoscedasticity must the residuals of the variances remain the same when moving along the best fitting line. The assumption is hetroscedastic when there appears to be a pattern in the residuals of the variables. In order to test for
homoscedasticity are scatterplots being used. Because there are four dependent variables and three years are being tested amount the total amount of best fitting lines twelve, but in order to provide a comprehensive overview only the scatterplots for ROA are presented below in figure 1, figure 2 and figure 3. The assumption for homoscedasticity for our regression model is being satisfied.
29
Figure 1: Scatterplot residuals ROA 2014 Figure 2: Scatterplot residuals ROA 2015
Figure 3: Scatterplot residual ROA 2016
3. Linear relationship
The third assumption is linear relationship; this means that there is a linear relationship between the dependent variables and independent variables. Again is this tested by making use of the scatterplots and visualize the relationship between the dependent and
independent variables. Figure 1, Figure 2 and figure 3 provide evidence that this assumption is also being satisfied.
4. No auto-correlation
The fourth assumption is that there must be no auto-correlation in the model, this can occur when data form the current period is affecting the data from other periods. To test for autocorrelation the Durbin-Watson test is used and provides values between 0 and 4 which are called the d-values. The four dependent variables during the period 2014 to 2016 are
30
provided in the table below. the general rule for the d-values to have no auto-correlation is that the values need to be between 1.5 and 2.5.
Durbin-Watson d-values 2014 2015 2016 ROA 2,088 1,835 1,813 ROE 2,000 1,880 1,976 FCF 1,667 2,003 1,997 AS_GROWTH 2,023 0,607 0,032
Table 10: Durbin-Watson d-values
Table 10 provides no evidence for auto-correlation for ROA, ROE and FCF. AS_GROWTH appears to be an exception for the years 2015 and 2016. In general the assumption for having no auto-correlation is satisfied.
5. No multicollinearity
The fifth and final assumption is multicollinearity, there must exist none or little
multicollinearity between two independent variables. This exists when two independent variables are highly correlated, the threshold for the Pearson correlation matrix is when above 0.9. Because the independent variables are not used in the same year in one model will three correlation matrixes asses if there is multicollinearity in this model.
Pearson's Correlations Matrices
WA_SCORE DFAMILY REV
CSP_WA Pearson Correlation 1 ,001 ,265 DFAMILY Pearson Correlation ,001 1 ,158 SIZE Pearson Correlation ,265 ,158 1
31
WA_SCORE DFAMILY REV
CSP_WA Pearson Correlation 1 ,063 ,334 DFAMILY Pearson Correlation ,063 1 ,168 SIZE Pearson Correlation ,334 ,168 1
Table 12: Pearson’s Correlation Matrix 2014
WA_SCORE DFAMILY REV
CSP_WA Pearson Correlation 1 ,086 ,263 DFAMILY Pearson Correlation ,086 1 ,209 SIZE Pearson Correlation ,263 ,209 1
Table 13: Pearson’s Correlation Matrix 2014
Table 11, table 12 and table 13 show that none of the independent variables has a Pearson correlation score above 0.9, thus there appears to be no problems with multicollinearity for the models in 2014, 2015 and 2016
In short the five assumptions mentioned in section 5.3 which are needed for this regression are being satisfied and therefore can the results from the regression be taken reliably.
5.4 Regression results
In the following chapter the results from the regression are being presented and early indications on the second hypothesis are being given. The regression is analyzing the second set of models which was; FP i;t = α1 + β1 CSP_WA i;t + β2 DFAMILY i;t + β3 SIZE i;t + ε. Where FP i;t is the financial performance measures ROA, ROE, FCF and AS_GROWTH. The independent
variables are CSP_WA, DFAMILY and SIZE where DFAMILY is a dummy variable which
contains a 1 for being a family firm taken from the GFBI and is SIZE a controlling variable for firms size in terms of revenue. The data sample is taken from the ASSET4 ESG database from Thomson Reuters and contains a total amount of 874, 1483 and 1002 companies
respectively in 2014, 2015 and 2016 which also respectively included 33, 37 and 26 family firms during those years.
Coefficientsa ROA 2014 ROA 2015 ROA 2016
B Std. Error Sig. B Std. Error Sig. B Std. Error Sig.
32
1 (Constant) 4,692 ,824 ,000 2,005 ,548 ,000 3,584 ,510 ,000
CSP_WA ,019 ,014 ,160 ,038 ,011 ,000 ,027 ,012 ,022
DFAMILY -2,038 1,966 ,300 2,359 2,131 ,269 2,398 2,020 ,235
SIZE 5,563E-09 ,000 ,666 2,769E-09 ,000 ,854 1,617E-09 ,000 ,913
Significance model ,304b ,001b ,050b
R-square ,005 ,011 ,008
a. Dependent Variable: ROA
b. Predictors: (Constant), SIZE, DFAMILY, CSP_WA Table 14: Results of the regression for ROA in 2014, 2015 and 2016
Table 14 presents the results from the first regression which analyzes the following model: ROA i;t = α1 + β1 CSP_WA i;t + β2 DFAMILY i;t + β3 SIZE i;t + ε
The R-squares in the model represents the variance in the dependable variable which is caused by the independent variables. With R-squares of 0.5% 1.1% and 0.8% for respectively 2014, 2015 and 2016 are these values fairly low which means that low part of the variance in ROA is been caused by CSP_WA, DFAMILY and SIZE. The ‘B’ in table 14 provides information on how the independent variables are relating to the dependent variable. The B for CSP_WA shows that there exists a positive relation for every year. Though DFAMILY shows a negative relation in 2014 but shows a positive relation in 2015 and 2016. Furthermore isn’t any independent variable significant for 2014 and DFAMILY doesn’t indicate any significant relationship with a 5%-level interval for 2015 and 2016 ether. Contrary shows CSP_WA a significant relationship for 2015 and 2016, indicating that having a higher score on corporate sustainability performance is positively related with the return of assets for companies in the tested population. But being a family firm does not seem to influence the ROA significantly based on these results.
Coefficientsa ROE 2014 ROE 2015 ROE 2016 B Std. Error Sig. B Std. Error Sig. B Std. Error Sig. 1 (Constant) -27,512 19,440 ,157 -,133 3,264 ,967 -8,425 7,635 ,270 CSP_WA ,617 ,319 ,054 ,285 ,064 ,000 ,385 ,175 ,028 DFAMILY -,260 46,382 ,996 1,140 12,680 ,928 5,531 30,261 ,855
SIZE -3,456E-09 ,000 ,991 2,362E-08 ,000 ,792 6,065E-08 ,000 ,785
Significance model ,261b ,000b ,125b
R-square
,005 ,016 ,006
a. Dependent Variable: ROE
b. Predictors: (Constant), SIZE, DFAMILY, CSP_WA Table 15: Results of the regression for ROE in 2014, 2015 and 2016
33
Table 15 relates to the second set of models which is analyzing dependent variable ROE: ROE i;t = α1 + β1 CSP_WA i;t + β2 DFAMILY i;t + β3 SIZE i;t + ε
The R-square values indicate that 0.5%, 1.6% and 0.6% for respectively 2014, 2015 and 2016 are the independent variables causing the variance of the dependent variable in the model. Furthermore are the values for ‘B’ in table 15 positive for CSP_WA in every year and also for the variable DFAMILY are the values positive except for 2014. For CSP_WA 2015 and 2016 prove to be significant for a significance level of 5%. CSP_WA significance level is close to the significance threshold but barely isn’t this indicates that the importance of corporate sustainability performance is rising over the course of years and provides supporting evidence that CSP_WA and financial performance in terms of ROE are positively related. Thus first performing better on corporate sustainability performance also perform better in the financial performance measured in return on equity. Again the values for DFAMILY are not significant in any of the three years. This indicates that there is no relationship between being a family firm and their return on equity.
Coefficientsa FCF 2014 FCF 2015 FCF 2016 B Std. Error Sig. B Std. Error Sig. B Std. Error Sig. 1 (Constant) 23,586 2,741 ,000 -514,258 247,212 ,038 -765,803 538,987 ,156 CSP_WA -,046 ,045 ,304 7,172 4,846 ,139 11,846 12,355 ,338 DFAMILY -5,684 6,541 ,385 131,918 960,548 ,891 180,525 2136,152 ,933
SIZE -5,042E-08 ,000 ,240 -1,476E-07 ,000 ,983 8,806E-08 ,000 ,996
Significance model ,214b ,478b ,798b
R-square ,005 ,002 ,001
a. Dependent Variable: FCF
b. Predictors: (Constant), SIZE, DFAMILY, CSP_WA Table 16: Results of the regression for FCF in 2014, 2015 and 2016
Table 16 presents the results from the third regression which analyzes the following model: FCF i;t = α1 + β1 CSP_WA i;t + β2 DFAMILY i;t + β3 SIZE i;t + ε
The R-squares in the model have a value of 0.5%, 0.2% and 0.1% for respectively 2014, 2015 and 2016 which are again fairly low which means that low part of the variance in FCF is been caused by CSP_WA, DFAMILY and SIZE. Moreover the ‘B’ columns in table 16 show a
negative relation in 2014 between the independent variables; CSP_WA, DFAMILY, SIZE and the dependent variable FCF. In 2015 and 2016 the independent variables show a positive relation which is the opposite as 2014 but is similar as the other financial performance
