Agricultural succession under pressure : an exploratory study of the successfactors of agricultural succession

(1)

Agricultural Succession

Under Pressure

An exploratory study of the successfactors of

agricultural succession.

Student: Ruud Paauwe Student number: 10475567

Date final version: 10th January 2015 Supervisor: Dr. Ir. Jeroen Kraaijenbrink Amsterdam Business School

University of Amsterdam

Faculty of Economics and Business Master Business Studies

(2)

1

Abstract

The number of agricultural companies is decreasing since the 1960’s, caused by a decreasing overall profitability and decrease in successions. Therefore, the aim of this study was to get a better understanding of the successfactors of an agricultural succession in order to provide successors with a framework of which competencies they could improve to increase their chances of a successful succession. A succession was determined to be successful based on the performance of the firm and the satisfaction of the successor. Linear regression was used to assess the influence of the flexibility, creativity and environmental awareness of the successor on the chances of a successful succession. Furthermore, the influence of training on these three competencies was investigated. The competencies were measured among 284 Dutch agricultural successors and 57 successions were evaluated on their successfulness. Results showed that the three competencies did not have a significant effect on the chances of successful succession. Surprisingly, experience, which was used as a control variable, did have a negative and significant effect on the chances of successful succession. Respondents who underwent training appeared to have higher levels of creativity and environmental awareness. The effect of overestimation of the competencies was not found to have a significant moderating effect on either the relationship between training and the competencies, nor did it have a significant moderating effect on the relationship between the competencies and successful succession. While most investigated effects appeared to be not significant, this study had some important implications for current research and practise. Based on the findings in this study, a recommendation can be made for future successors to work in a different company before the succession takes place.

(3)

2 Table of Contents Introduction ... 3 Literature Review ... 5 Succession ... 5 Competencies ... 8 Environmental Awareness ... 10 Creativity ... 10 Flexibility ... 11 Training ... 12 Self-awareness ... 13 Conceptual Model ... 15 Methods ... 16 Research Design ... 16

Variables and Measures ... 17

Sample and Data Collection ... 18

Statistical procedure... 20 Results ... 22 Reliability Analysis ... 22 Correlation Analysis ... 24 Regression Analysis ... 27 Mediation analysis ... 31 Moderation analysis ... 33 Discussion ... 35 Summary of Findings ... 35

Implications for Research ... 36

Implications for Practice ... 37

Limitations and Recommendations for Future Research ... 38

Conclusion ... 40

References ... 42

Appendix A: Items used for Variable Construction ... 46

(4)

3

Introduction

Never before was there such a huge growth in the world population as there is today. The world population is bursting at the seams, and with it the demand for food. The agricultural sector plays a key role in the worldwide food supply, but agricultural companies are slowly disappearing. Within the last decade the number of agricultural companies in the Netherlands alone decreased by almost 30.000 since the start of the new millennium and in 2014 approximately 67.000 agricultural companies are still operating. (Centraal Bureau voor de Statistiek, 2014).

According to a study from the Agricultural Economic Institute of the Wageningen University (Landbouw Economisch Instituut (LEI)) this current decline in the number of agricultural companies of 2-3% on average per year will continue until at least 2020 (Silvis, et al., 2009). They attributed part of this decline to the decreasing overall profitability in the agricultural sector, which results in less companies being able to compete. However, they also concluded that succession in the agricultural sector is declining and this will also lead to far fewer agricultural companies in 2020. The average age of farmers is already 52 years and is rising rapidly, with only 3.6% of the farmers in Europe under the age of 35.

The declining number of successions in the agricultural sector has been a concern as early as the 1960s (Clawson, 1963). Many factors have been researched but most of the researched factors don’t have a sufficient explanatory value on their own which is mainly attributed to lack of knowledge of other, unobservable factors (Hennessey & Rehmann, 2007). According to Gale (2003), there are two main reasons for the decline in succession in the agricultural sector. The first reason is that there are entry barriers which prevent successors from entering the agricultural sector. These include the high and increasing capital requirements and lack of financing for a succession. The second reason is that many successors have better earnings prospects in nonfarm careers, making it less appealing to succeed in the agricultural sector. Most researched financial causes for the decline in succession all fall within one of the two reasons suggested by Gale, and provide more detailed information why these reasons are valid. Farm size (Barcley, Foskey, & Reeve, 2005), profitability (Hennessey & Rehmann, 2007) and total farm assets (Calus, Huylenbroeck, & Lierde, 2008) are merely some of the researched causes in the literature, which are a more detailed explanation of the suggested reasons of Gale.

(5)

4

stop this decline is to improve the profitability in the agricultural sector. Indeed it has been proven that an increase in the received subsidy payments, output prices and off-farm income increase the profitability and reduce the decline in the number of agricultural companies (Breustedt & Glauben, 2007). However, these solutions all focus on an organizational level, instead of an individual level. By focusing on an individual level, successors could be able to change or improve themselves before the succession takes places, instead of after. Indeed it seems that higher formal education of the successor increases the chances of having a successful succession (Kimhi & Nachlieli, 2001). In an explanatory study Chrisman et al (1998) investigated 29 different personal skills and attributes of a possible successor that the incumbent considers important. However, that study focused on what the incumbent considers important skills and attributes for a successor, not whether these skills and attributes actually lead to a successful succession. Secondly, this study focused on all family-owned businesses, regardless of which sector the business operates in.

This study will therefore focus on which personal skills and attributes will make a succession in the agricultural sector actually successful in practice. The purpose is to get a better understanding of what makes an agricultural succession successful, increase the chances of a successful succession and help possible successors determine if they are suitable to success. The conclusions will be solely based on the agricultural sector, because current researches are based on FOB (family owned businesses) in the small and medium-sized enterprises (SMEs).

(6)

5

Literature Review

This chapter discusses the most relevant findings from the current literature about succession and states the hypotheses of this research. Initially, it starts with a description of what succession is and what makes it successful. Subsequently, this chapter continues with a discussion of the competencies that might play a role in successful succession. In addition, this chapter outlines how training these competencies might increase the chances of successful succession. Finally, the influence of overestimation of a successors competencies is discussed.

Succession

In order to study what makes a succession successful it is necessary to begin by stating what succession is. The definition used by the Dutch Tax and Customs Administrations is as follows: the acquisition and continuation of a company, or part of, by donation or inheritance. This definition is, however, too specific, since not all successions are the result of inheritance or donation. In general, succession is seen as the introduction of a new individual (or successor) within an already existing company and the exit of the already present entrepreneur, resulting in a shift of possession and change of the status of the introduced individual from non-entrepreneur to entrepreneur. The way the shift in possession is accomplished is not relevant. This will also be the definition used in this research; succession is a transition from being a non-entrepreneur to an entrepreneur as a result of the shift in possession of the already existing company.

Although succession in its essence can be seen as a simple transition, the process that leads up to the succession is far more complex. An important part of this process consists of personal and professional development of the successor. (Kesner and Sebora, 1994). The Entrepreneurs Academy classified three different phases in this development, which are based on a theory developed by Lievegoed (1969). In the first phase of development, the focus lies primarily on the skills and tasks required for a profession in the agricultural sector. In this phase the future successor can be considered more as an employee or craftsman than an entrepreneur. The professional development is followed by the management development phase. In this phase the future successor develops himself on an organizational level, with a focus on leadership and management. The last phase of

(7)

6

development is the entrepreneurial development, in which the successor develops into a mature entrepreneur.

Now that the definition of succession and the process leading up to it is explained, it is necessary to determine what successful succession is. Multiple attempts have been made to identify successful and unsuccessful successions. There is, however, little consensus on what constitutes a successful or effective succession. Early research suggested that the satisfaction of the incumbent, the successor and other family members with the succession process can be used as an good assessment of the success (Handler, 1989; Sharma, 1997; Stempler, 1988). Others did, however, point out that not only the satisfaction of stakeholders with the succession process defines a successful succession, but also the successor’s ability to keep the family business healthy (Harvey & Evans, 1995; Hume, 1999 ; Goldberg, 1996). Le-Breton-Miller, Miller and Steier (2004) concluded in their research that most used definitions of successful succession are either based on the subsequent positive performance of the firm and the ultimate viability of the business, or the satisfaction of stakeholders with the succession process.

It has also been argued that the process of succession should be considered more from an entrepreneurial perspective; entrepreneurial success should form the foundation of

(8)

7

what the definition of successful succession is (Habbershon & Pistrui, 2002; Nordqvist & Melin, 2010). This makes sense, because succession is in its essence a transition from non-entrepreneur to non-entrepreneur and thus the success of a succession depends heavily on the success of the entrepreneur. Therefore, this study uses entrepreneurial success as a guideline to determine a definition for successful succession.

The current literature on entrepreneurial success is extensive, which is caused by (again) a lack of consensus on what a successful entrepreneur is and therefore how it should be measured (Sørensen & Chang, 2006). Performance measurements, depending on the definition used, include firm sales, firm revenue, sales growth and revenue growth, to name a few. However, when only performance is used to measure the successfulness of entrepreneurs an issue arises because there might be a discrepancy between the externally visible measures of entrepreneurial success and the internal measures of success held by the entrepreneur himself. Some entrepreneurs value non-monetary rewards, like autonomy and entrepreneurial identity, more than monetary rewards (Xu & Tuef, 2004) .

Since there are two methods to determine whether a succession was successful, a combination of performance and satisfaction will be used. The performance will be based on an improved or equal net profit one year after the succession in comparison to the average of the three years prior to the succession. As a result, fluctuations or incidental results in the net income before the succession will be compensated for. However, a comparison between the net profit of the successor and the incumbent might be skewed, due to the fact that the successor must finance his acquisition of the company and therefore has higher interest costs and depreciations than the incumbent. Therefore, the interest costs and depreciations will not be included in the net profit so that the differences in net-income between the successor and the incumbent are more comparable. The measure of the satisfaction with the succession will be based on a self-assessed rating of the successor. When performances are inconclusive on the successfulness of the succession, satisfaction will be the decisive factor.

(9)

8

Competencies

The literature about which factors influence a successful succession is extensive. Most researched factors are concerned with the interaction between the incumbent and the successor, according to a review of the literature by Le Briton-Miller et al (2004). They concluded that most factors concern the family relationships, willingness of the incumbent to hand-off power, shared vision, and succession planning. While these factors are capable of predicting the satisfaction of stakeholders with the succession, they are not capable of determining whether the successor will be a successful entrepreneur with a matching high performance of the company. Since this study measures the success of a succession based upon the subsequent performance of the firm, the successor should possess the qualities that are needed to be an successful entrepreneur.

Indeed, business financiers find the personal characteristics of an entrepreneur to be an important predictor of entrepreneurial success (Shepherd, 1999; Zopounidis, 1994). However, the literature about a successors characteristics in relation to high performance is limited. A study by Barach and Ganitsky (1995) suggested that certain characteristics improve the credibility with the incumbent. Another study by Chrisman et al (1998) consisted of a ranking of the characteristics the incumbent considers important. The qualities of a successor that affect the subsequent performance after the succession are relatively less investigated.

To differentiate between successor’s individual qualities, the theoretical framework developed by Man, Lau and Chan (2002) is used in this study. In this framework, the influence of entrepreneurs on the performance of the company is considered critical and is addressed through the competency approach. Competencies offer a broader spectrum than qualities, since they are formed by a combination of skills, abilities, knowledge and characteristics that are associated with a high performance on a job (Mirabile, 1997). Competencies are thus constructs and therefore involves inferring the existence of a concept that is not directly measurable or observable based upon related information (Pedhazur & Schmelkin, 1991; Nunnally & Bernstein, 1994). According to Bird (1995) competencies are behavioural and observable, and are thus more closely linked to performance than other entrepreneurial characteristics, such as personality traits, intentions or motivation.

Man, Lau and Chan (2002) categorised six major areas of entrepreneurial competencies, consisting of opportunity, relationship, conceptual, organizing, strategic and

(10)

9

commitment competencies. These areas are based on the assumption that each area affects a company’s performance in different ways though direct and indirect effects. Lance et al (2005) used this framework to measure each area with Dutch agricultural entrepreneurs. They concluded, based on self-assessed ratings of the entrepreneurs themselves and ratings of peers and experts, that Dutch agricultural entrepreneurs rated and scored the lowest on the opportunity competence area. The opportunity area consists of the competencies responsible for the discovery and development of new entrepreneurial opportunities. Since the success of the succession is based on the subsequent performance after the succession, successors who possess the competencies in the opportunity area could have a competitive advantage.

According to Shane and Venkataraman (2000) the exploration and development of new markets is the absolute key to entrepreneurial success, because “to have entrepreneurship, you must first have entrepreneurial opportunities”. The findings of Lance et al (2005) are therefore remarkable, but consistent with a correlation analysis done by the Rabobank on production amount and prices in the Dutch agricultural sector. The Rabobank concluded that there is not enough demand for the produced amount of products in most agricultural branches, and entrepreneurs in should either collectively produce less, or find ways to create more demand. These two studies combined, suggest that an entrepreneur who is able to find and exploit new opportunities has a competitive advantage. Therefore, a successor who possesses the competencies needed for the discovery and exploitation of new business opportunities could have a higher chance of a successful succession, based on this competitive advantage.

Lance et al (2005) used the competencies general awareness, international orientation and market orientation as underlying competencies for the opportunity area. While these competencies are useful to detect and discover opportunities, they do not affect an entrepreneurs ability to actually exploit and develop these opportunities. Merely finding the opportunities does not provide a competitive advantage unless the opportunities are actually exploited. Therefore, this study uses environmental awareness as a competency to discover new entrepreneurial opportunities, and creativity and flexibility to develop and adapt to these opportunities. These competencies are chosen because successors could also possess these competencies, while other competencies as vision, strategic planning and innovativeness are all mostly developed when the successor actually succeeded and is an

(11)

10

entrepreneur. Furthermore, the literature shows that these competencies are all positively related to the discovery and exploitation of new opportunities and thus a higher chance to improve a company’s performance through the competitive advantage.

Environmental Awareness

Environmental awareness allows entrepreneurs to constantly scan their environment and to be persistent and unconventional in their attempts to investigate new ideas (Busenitz, 1996). It is, as Ray and Cardozo (1996) argued, a state of increased alertness to information before the opportunity is actually recognized by the entrepreneur. It consists of the knowledge and network of the operating environment available to the successor on both macro and micro levels. On a macro level, environmental awareness covers the demand and supply of the market, both local and international, and the ability to sympathize with the needs of (potential) clients. On a micro level, environmental awareness covers the personal setting of the successor and consists of the needs and ambitions of family members and/or the incumbent. Therefore, a successor who is more environmental aware has a higher chance to discover new opportunities when he or she succeeded. Therefore to investigate this relation, the following hypothesis is proposed:

H1a: Environmental Awareness has a positive effect on the chances of a successful succession, so that greater environmental awareness leads to greater chances of successful succession.

Creativity

Creativity has been viewed as the construction of ideas or products which are new and potentially useful (Amabile, 1998). It enables the entrepreneur to take advantage of opportunities (Shalley, Zhou, & Oldham, 2004), which can result in a competitive advantage and provides the basis for innovation and business growth (Bilton, 2007). Since creativity allows an entrepreneur to take advantage of opportunities, it could lead to a better performance. Therefore, successors who are creative could have a higher chance of a successful succession, based on the assumption that they can develop opportunities better compared to successors who are not creative. This relation is proposed in the following hypothesis:

H1b: Creativity has a positive effect on the chances of a successful succession, so that greater creativity leads to greater chances of successful succession.

(12)

11 Flexibility

Flexibility is the measurement of the ability to adapt to changes in the market and implementing new strategies. Adaptability is important for an entrepreneur when business environments are characterized by rapid, substantial, and discontinuous change (Hitt, 2000) which is the case for the agricultural sector. To realize and sustain a competitive advantage in such an environment, entrepreneurs must be able respond strategically and fast to changes in the environment (Ireland & Hitt, 1999). Therefore, successors who are flexible and thus better at responding and adapting to changing environments, are better at developing the discovered opportunities, thereby having a competitive advantage in comparison to those who are not flexible. Flexibility could thus increase the chances of a successful succession based on this competitive advantage, compared to successors who are not flexible. To investigate this relation, the following hypothesis is proposed:

H1c: Flexibility has a positive effect on the chances of a successful succession, so that greater flexibility leads to greater chances of successful succession.

(13)

12

Training

Policymakers believe that entrepreneurial education leads to increased levels of entrepreneurship (European Commission, 2006). For this to be true, entrepreneurial skills should improve when trained and should not be fixed personal characteristics. Indeed it has been shown that entrepreneurial skills can be improved by training itself (Nyström, 1979; (Sluis, Praag, & Witteloostuijm, 2006). However, there is limited research about the training of flexibility, creativity, environmental awareness and self-knowledge.

The research by Oosterbeek et al (2009) comes closest to finding whether training improves these skills. They showed that the training by the Junior Achievement Young Enterprise student mini-company (SMC) did not improve flexibility, creativity and market awareness. By using a difference-in-difference method, they compared the self-assessed ratings of participant’s skills. A pre-treatment group was compared to a control group and it showed no significant differences in the ratings between the two groups. But after the treatment, the ratings of the control group were actually higher than the treatment group. Oosterbeek et al attributed this to the fact that participants of the training obtained more realistic perspectives on themselves and as to what it takes to be an entrepreneur.

While it seems that the training did not actually improve flexibility, creativity and market awareness of participants, this study will use training as a possible successfactor of succession. There are a couple of reasons for this. First of all, the research by Oosterbeek et al only measures the self-assessed ratings. The decline in ratings after the training is attributed to more realistic perspectives, but that does not mean the ratings of the participants did not increase. Perhaps the participants valued their initial ratings to high, causing a decline when they obtained more realistic perspectives. If the participants had realistic perspectives from the start, their initial ratings might have been lower so that perhaps there indeed was an improvement. Secondly, the research by Oosterbeek et al 2009 included all possible starters from every market sector, while this research only focusses on the agricultural sector. Lastly, this research studies when a succession is successful. Whether the flexibility, creativity and environmental awareness of possible successors improve is not the only effect of training.

(14)

13

H2a: Successors who underwent training have higher levels of the competencies

flexibility, creativity and environmentally awareness compared to those who are not engaged in training.

H2b: The effect of training on the chances of a successful succession is mediated through the competencies flexibility, creativity and environmental awareness.

Self-awareness

Since this study measures the competencies through a self-assessed rating by the respondents, over or underestimation of the competencies could bias the results. Over or underestimation of the competencies is caused by low self-awareness, which is defined as the extent to which the self and other-raters agree on

the

level of the competencies the focal individual attains (Fletcher & Bailey, 2003). The literature on self-awareness of competencies is extended in large organizations and nonbusiness settings (Maurer, Wrenn, Pierce, Tross, & Collins, 2003). However, the literature on self-awareness in relation to entrepreneurs and successors of existing small businesses is limited. While this relation is promising, it is beyond the scope of this study. Self-awareness will be used as a control variable to more objectively investigate the relation between training and the competencies and the relation between the competencies and successful succession.

The necessity for to control the proposed relations comes from the fact that earlier research showed or suggested that Dutch agricultural entrepreneurs and successors have low self-awareness. To illustrate: research by Lace et al (2010) on the self-awareness of Dutch agricultural entrepreneurs showed that there is a tendency towards underestimation of the competencies. However, the earlier mentioned research by Oosterbeek et al (2009) suggested that successors tend to overestimate their competencies before they underwent training, causing a decreased effect of training on the competencies.

A discrepancy between the true or actual level of the competencies can moderate the proposed relations in two ways. Firstly, the effect of training on the level of competencies can be moderated through low self-awareness, whereby overestimation causes a decreased effect of training on the competencies as suggested by Oosterbeek et al (2009). Secondly, overestimation of the competencies could cause a decreased effect of the competencies on the chances of successful succession. This is because it is expected that the higher the level of the competencies are, the higher the chances of a successful succession.

(15)

14

Since overestimation leads to inflated levels of measured competencies, this proposed relation will be weaker. By controlling the proposed relations for overestimation, the results of this study will be more valid.

Therefore the following hypotheses can be proposed:

H3a: Overestimation of the competencies by a successor moderates the effect training has on the competencies flexibility, creativity and environmental awareness, so that the more a successor overestimates his or her own competencies, the lower the effect training has on the competencies.

H3b: Overestimation of the competencies by a successor moderates the effect of the competencies flexibility, creativity and environmental awareness on the chances of a successful succession, so that the more a successor overestimates his or her

competencies, the lower the effect the competencies have on the chances of successful succession.

(16)

15

Conceptual Model

From these hypotheses, the following model can be constructed. Flexibility, creativity and environmental awareness are competencies that might increase the chances of a successful succession. These competencies might also improve by training, making the competencies act as mediators in the relationship between training and successful succession. Self-knowledge is also important in this model, because the variables are constructed by a self-assessed rating of the participants. Over-estimation of participants own competencies could lead to an decreased influence of training on these competencies and also a decreased influence of the competencies on successful succession. Therefore, the variable overestimation is added to the model as a moderator. The direct relationships are indicated with bold arrows and big boxes, the moderation effect of self-knowledge with small boxes in figure 2.

Training Successful Succession

Flexibility Creativity Environmental Awareness Overestimation Overestimation

(17)

16

Methods

In this chapter the approach and design of this research will be explained. Initially, the research method will be discussed including strengths and limitations of the questionnaire method. Subsequently, this chapter describes how the variables will be measured and the used sample. Finally, this chapter ends with an explanation how the statistical procedure will be undertaken.

Research Design

The purpose of this thesis is to determine and quantify the success factors of an agricultural successful succession. Thereby, based on the current literature, the direct effects of flexibility, creativity and environmental awareness, are examined to see whether it influences the chances of a successful succession. Since training might also have a direct effect on flexibility, creativity, environmental awareness and therefore an indirect effect on the chances of a successful succession, the mediating role of those competencies is also examined.

This research has a quasi-experimental research design which is based on a natural experimental setting, consisting of a natural treatment group and a natural control group. The treatment group is formed by participants of the so-called Rabobank Successors Perspective (Rabobank Opvolgers Perspectief, ROP1) and other succession training programs which are aimed at helping future Dutch agricultural successors with their succession. The control group is formed by successors who did not participate in a training programme before the succeeded. This research is based on reported data of agricultural successors who already completed their succession and future successors, without any form of intervention of the data.

The temporal orientation of the data is cross-section; the valuation of the skills and successful succession is based on a single point in time. This means this research will not use a comparison between a pre-treatment and after-treatment group. The choice for this

1_{The ROP-program was launched in 2007 and consists of three parts. It starts with a 1.5 hour so called}

mirror-interview, performed by an external physiological agency. In this interview the future successor is asked questions on an individual level. The key questions are who am I, what kind of business do I want and what are my strengths and weaknesses. This leads to a discussion about the personal- and entrepreneurial skills of the possible successor and at what points he or she needs to develop. After that the Rabobank provides a financial scan in which participants can calculate themselves whether the acquisition is financially feasible. Finally there is a 5-day Successor Training where participants write their own business plan for their succession and are trained in entrepreneurial competencies, including flexibility, creativity and environmental awareness.

(18)

17

method is based on the fact that the population who participated in the ROP-program is limited. From 2007-2014 only 650 participants followed the entire ROP program. If a longitudinal research was used, the pre-treatment group would be minimal and therefore the research would look more like a one-shot case study, which in turn leads to a low validity of the conclusions.

Variables and Measures

Successful succession

To measure whether a succession was successful, the respondents were asked to estimate the development of three key financial indicators one year after the succession was completed. A five-point Likert scale ranging from 1 (strong decreased) to 5 (strong increased) was used. The key indicators were: (1) net income, (2) total depreciation and (3) interest costs. In addition, the respondents were also asked to rate their satisfaction with the succession on a five-point scale, with five being the highest and 1 the lowest. The average of the four items combined forms a continuous variable, which indicates the success of the succession.

Competencies

The competencies are measured by self-assessment of the respondents, using a 5 point Likert-scale ranging from 1 (strongly disagree) to 5 (strongly agree). The items used in the questionnaire are derived from the Entrepreneur Scan (E-Scan), as was developed by Driessen and Zwarts (2005). The items of the E-Scan are based on different, individual competences, where each competence consists of four components: (1) knowledge and experience, (2) motivation, (3) capabilities and (4) characteristics. The E-scan is validated by the University of Groningen and over 500.000 entrepreneurs already did the test. Some questions were also reverse coded to reduce the acquiescence bias (known as yea-saying) and extreme response bias (Knowles, 1997; Podsakoff, MacKenzie, & Podsakoff, 2003). Eleven items were used to assess a respondent’s flexibility, eight for creativity, and fourteen for environmental awareness. The respondents average score in each of these competencies will be used for statistical procedure, whereby the rating of the negatively phrased questions will be inverted. However, not all items will be used for the regression analysis but only the items that maximize Cronbachs alpha of each competency. This will be addressed in the Reliability Analysis in the Results section.

(19)

18 Overestimation

To determine if a respondent overestimated his or her competencies, a combination of six items was used. Based on a 5 point Likert scale, the respondents had to rate themselves on three items. Additionally, those same three items were rephrased in a way that respondents had to rate themselves on the same items but this time based on the perspective of others. When the respondents rated themselves higher than others would, it indicated that the respondents overestimated themselves. To create a variable from these items, the respondents’ ratings of themselves is compared to the ratings of the corresponding item from the perspective of others. The difference in rating by the respondent on the corresponding items receive a score, ranging from 1 (self-assessed rating is equal or lower than the rating from the perspective of others) to 5.

Control variables

Results of the current literature are controlled for Gender, Age and Experience, whereby experience is measured as the years the respondent is active in the agricultural sector. These items were included in the first section of the questionnaire.

Sample and Data Collection

An online questionnaire was used to collect the primary data, because it allows collecting a large amount of data in relatively short time period. Furthermore, it is the least invasive way to collect personal information and it is the least labour-intensive method, both for the interviewer as well as the respondent.

The sample consisted of Dutch agricultural successors, both those who have completed their succession as well as those that are planning to do so in the future. The participants were approached both directly and indirectly; indirectly by NAJK2, who spread the questionnaire among their members, and directly when they participated in the ROP-program and their contact details were known at Rabobank. In total, 155 participants were approached directly and around 8,000 from the NAJK. This makes the response rate more difficult to determine since it is unknown how many people have seen the invitation exactly. When it is assumed that 8,200 people have received the invitation, the response rate is 3.7% with a total of 299 completed questionnaires (in total 15 respondents were removed from

2_{NAJK stands for Nederlands Agrarisch Jongeren Kontakt or the Dutch Agriculture Youth Contact. It is a}

association for agricultural youth (under 36 years) in the Netherlands. NAJK has about 8,000 members and is active in local, provincial, national and European level. The association is committed to the interests of young farmers (up to 35 years) in the Netherlands.

(20)

19

the sample, due to their lack of qualification to participate in this study and little variance in their individual answers to the questionnaire). This percentage is low , despite the fact that a cadastre ruler is promised as an incentive to complete the questionnaire. However, a similar study (family owned business research) that also used the members of NAJK as sample still had an even worse response rate of 185. International there is a concern about the non-response in surveys; it depends strongly on the fieldwork and country (De Leeuw, E.D & Hox, J,J. 1998). It cannot be assumed that this is also the case in the Dutch agricultural sector. The participants were on average 30 years old, and only 39 (13,4 %) of respondents were women. The participants who respondent in the first week of admission were on average 30 years old, while respondents in the last week were on average 32 years old. Of the respondents who reported to have participated in the ROP-program (15,7%) or similar training or courses 19,1%), 22,7% also answered the items about their succession. A significant increase in the number of respondents who underwent training was found in week 3 and 4 of admission (58% and 61%) in comparison to the first two weeks (34% and 19%). This difference can be attributed to the fact that at the end of week 2, known participants of the ROP-program were directly contacted.

Participants scored on average 4.0597 on flexibility, 3.6318 on creativity and 3.3401 on environmental awareness. The score of the competencies was surprisingly consistent between early and late respondents. Participants who respond in the first week scored on average 4.0669 on flexibility, 3.6348 on creativity and 3.3374 on environmental awareness. In comparison: participants who respond in the fourth week scored on average 3.9409 on flexibility, 3.6398 on creativity and 3.328 on environmental awareness. An ANOVA analysis revealed that these differences were not significant, with a significance of p=0.242 for flexibility, p=0.999 for creativity and 0.978 for environmental awareness.

To test the moderating roles of the independent variables, only a small portion of the total sample was used, as it is necessary that the participants already completed their succession for those. In total, 57 respondents, who reported to have completed their succession, answered the financial questions about their succession. Of the 57 successors who have answered the financial questions, creativity had the largest variance (0.36), and flexibility had the smallest variance (0.2304). This does not seem large, but when the highest value for a competency is 5 and the lowest 1 this is still a considerable variance, which

(21)

20

suggests that the questionnaire was distinctive between respondents. To examine the mediating role of the competencies, the complete sample was used.

Statistical procedure

Survey administration started on October 20th 2014 and was closed four weeks later on November 18th 2014. The statistical analyses were performed, using the Statistical software Package for Social Sciences (SPSS). The statistical analyses will consist of two distinct parts: (1) the effect of training on flexibility, creativity and environmental awareness and (2) the effects of flexibility, creativity and environmental awareness on the chances of a successful succession. The first part will be referred to as Model 1 and the second part as Model 2, whereby both models will be analysed through linear regression. This distinction between the two models is important because Model 1 will be constructed with the data of the complete sample (n=299). As it turns out, in total 15 respondents were removed from the sample, due to their lack of qualification to participate in this study and little variance in their individual answers to the questionnaire. After this correction the complete sample is n= 284, whereas Model 2 will be constructed with a subsample (n=57). The reason for this is that 57 respondents answered the items about their succession they actually realized their acquisition, these answers of the subsample group are vital to the evaluation of the succession. It would, however, be a wasted of data when only the 57 responses were used to test the direct effect of training on the competencies, when 284 are available.

The statistical procedure will initially begin with a reliability analysis of the measures and the sample. Only the items that maximize the internal consistency (Cronbach’s alpha) are used to construct the variables with the measurements of the questionnaire. According to Pallant (2007) an internal consistency of 0,7 is acceptable. Furthermore, all respondents’ answers are checked for eligibility to be eliminated from the data. Reasons for elimination could be incomplete answered questionnaires, lack of understanding the instructions, no variance in an individual answers (e.g. all multiple choices will have the same answer) or lack of qualification to participate in the questionnaire.

Subsequently, a correlation analysis is undertaken to determine whether there is multicollinearity between the independent variables of both models. Multicollinearity occurs when two or more dependent variables are correlated, meaning that one can be predicted from the other variables. A simple way of detecting multicollinearity between the variables is

(22)

21

to create a correlation coefficient matrix of all the variables, whereby strong and significant correlations indicate that there might be multicollinearity. The most commonly used correlation coefficient is the Pearson correlation coefficient. When a strong and significant correlation is shown in the correlation matrix, it could indicate multicollinearity. To test for multicollinearity, the tolerance and variance inflation factor (VIF) are used. The tolerance of a particular variable is 1 minus the R2 that results from the regression if the other variables predict that particular variable. The corresponding VIF is derived from the tolerance, and is simply 1 divided through the tolerance. If all the variables are completely uncorrelated, both the tolerance and VIF are 1 for each variable. If the variables are closely related, the tolerance goes to 0 and the VIF gets large.

Additionally, the direct relationships in both Model 1 and Model 2 are tested with linear regression. The variables of Model 2 will be selected by backward elimination. This involves the process of starting with all variables and deleting variables that improve the model the most by being deleted, step by step. The reason for this is that more variables are used in Model 2, and this could lead to ‘overfitting’ of the model. When overfitting occurs, all independent variables will not be significant, but the overall model will be significant. To test the mediating role of the competencies, an SPSS macro of Preacher & Hayes (2008) was used. With this macro, the Sobel-test can be performed which includes the mediator (training) to the model. When the effect of the competencies is reduced and the effect of the mediator remains significant, it suggests that there is a mediating effect of training. Finally, the moderating role of overestimation will be tested by adding interaction terms of overestimation to both models using model 1 of Process. When the interaction terms prove to be significant, it suggests that overestimation acts as a moderator.

(23)

22

Results

In this section the results of this research are reported. The analysis initially begins with the reliability of the measured variables. Subsequently, a correlation matrix of the variables is constructed. Finally, to discuss the direct relationships, linear regression is performed, in advance of the mediation and interaction effects respectively.

Reliability Analysis

Reliability is a precondition of validity. As was already mentioned before, this research had a quantitative approach and therefore a questionnaire was used in order to measure the variables. To assure the reliability of these measurements, the data is checked whether there are reasons for elimination from the sample. The complete sample that is used for regression analysis is n=284 (Model 1), and the subsample to n=57 ( the successors that already (Model 2).

Subsequently, the variables that are used for regression analyses are constructed by maximizing Cronbach’s alpha for each variable. This is done by selecting the items that maximise the internal consistency, so that the items that are used to construct the variables for flexibility, creativity and environmental awareness are all consistent. The internal consistency is measured by calculating Cronbachs alpha, whereby an internal consistency of 0.7 is good and acceptable (Pallant, 2010). An overview of the items that are used for regression analyses is presented in Appendix A.

The maximised internal consistency for the variable flexibility consists of six items (before 11 items) and has an alpha of 0.757 for the complete sample and 0.663 for the subsample. Since the alpha of the complete sample is higher than the alpha of the subsample, it suggests that the alpha of flexibility of the subsample would have reached the acceptable level of 0,7 if the number of respondents were higher. The alphas show that 75.7% and 66.3% of the variability of flexibility in both samples can be considered as the true score variance of reliable variance. Thereby, no negative values were found in the inter-item correlations of the complete sample. It can be concluded that the six items were always measuring the same underlying characteristic in the complete sample and to a lesser extent of the subsample.

The maximised Cronbach’s alpha for the variable creativity consists of 6 items (before 8) and has an alpha of 0.785 for the complete sample and 0.827 for the subsample. Therefore, 78.5% of the variability in the composite score of creativity is considered as the

(24)

23

true score variance of reliable variance of the complete sample, and 82,7% of the subsample. Internal consistency of both measurements is assured, since both alphas are higher than 0.7 (Pallant, 2007). Also, the inter- item correlation matrix showed no negative values, indicating that the six items were always measuring the same underlying characteristic.

The maximised alphas for environmental awareness are 0.773 and 0.8 for the complete and subsample. Therefore, 77.3% of the variability in the composite score of creativity is considered as the true score variance of reliable variance in the complete sample, and 80% of the subsample. Internal consistency is assured, because both alphas are higher than 0.7 (Pallant, 2007). The inter- item correlation matrix showed no negative values, so the six (before 14) items were always measuring the same underlying characteristic.

The variable overestimation is measured using 6 items. Cronbach’s alpha for this variable is 0.512 for the complete sample and 0.597 for the subsample. These values are considerably lower than the acceptable level of 0.7. However, these questions come from the validated E-scan questionnaire. Furthermore, a low alpha indicates low inter-tem correlations, which could be expected when measuring if the respondent assesses himself the same as others. A low alpha could mean that others assess the respondent substantially different than the respondent himself. Therefore, despite the low value of the alphas, the six items which measure overestimation will still be used in this research.

The variable successful succession is constructed from 4 items, which have an internal consistency of 0.296. This low value could be expected, since the items that construct the variable successful succession all measure a different factor, in which each factor is distinctive. To illustrate: the satisfaction a respondent has with his or her succession measures a complete different factor than the total appreciations and thus have a low consistency. Hence, although the internal consistency between the four items measuring successful succession is low, it does not make the variable successful succession less reliable. Therefore the four items that measure successful succession will still be used in this research.

(25)

24

Correlation Analysis

Model 1

An overview of the descriptive statistics and correlations of Model 1 is presented in Table 1. A first observation is that creativity and environmental awareness are significantly and strongly correlated with training (r=0.167, p<0,01; r=0.192, p<0.01). It appears that respondents who underwent training have a higher level of creativity and environmental awareness, which is in correspondence with hypothesis 2a. More striking were the strong and significant correlations between the competencies themselves. Since Model 1 uses the competencies as dependent variables, multicollinearity is not an issue that has to be dealt with in this model. It does, however, suggest that these competencies are all part of the opportunity area of related competencies, as was described in the Literature Review. These three competencies could thus indeed be important for discovering new business opportunities.

Finally, experience is found to be strongly correlated with age (r=0.732, p<0.01), which is not surprising. It is, however, surprising that experience is also negatively correlated with gender (r=-.201, p<0.01). This could indicate that, since most agricultural companies are family owned, men start at a younger age to help with daily operations. This relation will not be further investigated, because it is beyond the scope of this research.

Table 1: Means, Standard Deviation and Correlations of Complete Sample (Model 1) Number of items Variables M SD 1 2 3 4 5 6 7 8 1. Gender 1 .14 .35 - 2. Age 1 29.67 7.12 -.071 - 3. Experience 1 12.13 7.48 -.202** .734** - 4. Training 1 .36 .48 .070 .151* .087 - 5. Overestimation 6 1.64 .43 -.105 .091 .108 .084 - 6. Flexibility 6 4.06 .47 .015 -.037 .012 .100 -.046 - 7. Creativity 6 3.63 .61 -.101 -.040 .045 .167** -.058 .433** - 8. Environmental Awareness 6 3.35 0.71 -.005 -.028 -.003 .192 ** -.054 .276** .407** - Note: N=284

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

(26)

25 Model 2

In Table 2, an overview of the descriptive statistics and correlations of Model 2 is presented, based on the data of the subsample. To start, there is a strong and significant correlation between environmental awareness and successful succession (r=.133, p<0.05). A relatively stronger correlation was found between the flexibility and successful succession (r=0.191, p<0.05). These two correlations support hypotheses 1a and 1c, that the more flexible and environmental aware a successor is, the higher the chances of a successful succession. Surprisingly, creativity was not found to be significantly correlated to the chances of a successful succession (r=0.083, p>0.05).

Subsequently, training is correlated with overestimation of the competencies (r=0.320, p<0.05), which indicate that respondents who underwent training overestimate their competencies more. This correlation supports the proposed moderator effect of overestimation (hypothesis 3a), but the effect is in a different direction than was expected. However, this correlation did not exist in the complete sample indicating that with a larger sample training and overestimation are not correlated.

Furthermore, experience and age are again highly correlated (r=0.743, p<0.05), as was the case with Model 1. While not surprising, it could result in the elimination of one of the variables if multicollinearity arises between the two.

Finally, the correlations between the competencies themselves are strongly correlated, and are even stronger than they were in Model 1. Flexibility is strongly correlated with creativity (r=0.517, p<0.05), creativity is strongly correlated with environmental awareness (r=0.509, p<0.05) and flexibility is strongly related with environmental awareness (r=0.589, p<0.05). Because the competencies are used as independent variables in Model 2, this strong correlation could indicate that there is multicollinearity between the variables. To test for multicollinearity, the tolerance and VIF values are presented in Table 3. According to O’Brien (2007) a tolerance less of 0,2 and a VIF higher than 5 indicate a multicollinearity problem.

Table 3 shows that the competencies have a high degree of multicollinearity, since the VIF values are way above 5, and the tolerance values are below 0.2. Since the three competencies all fall within the same range (1-5), a way to solve this problem is to mean-centre the competencies. Mean-centring involves the subtraction of the variables average, thus creating a new range where a positive value indicates an above average level of

(27)

26

competence, and a negative value of the variable an under average level of the competence.

These mean-centered competencies are presented in Table 4. Now, all three competencies have VIF values below 5 and tolerance values above 0.2. In the remaining of this research the competencies will be using this mean-centered value for Model 2.

Table 2 Means, Standard Deviation and Correlations of Subsample (Model 2) Number of items Variables M SD 1 2 3 4 5 6 7 8 9 1. Gender 1 .18 .38 - 2. Age 1 35.53 7.38 -.040 - 3. Experience 1 17.88 8.98 -.279* .733** - 4. Training 1 .32 .47 .084 -.142 -.046 - 5. Overestimation 6 1.69 .40 -.143 -.050 .014 .320** - 6. Flexibility 6 4.00 .49 .130 .019 .085 .134 -.126 - 7. Creativity 6 3.56 .69 .017 -.116 -.074 -.039 -.111 .517** - 8. Environmental Awareness 6 3.30 .77 -.103 -.028 .012 .018 .031 .509** .589** - 9. Successful Succession 4 3.70 .51 0.048 -.277* -.334* .258 .130 .191* .083 .133* - Note: N=57

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

Table 3: Multicollinearity diagnostics of Model 2

Variables Tolerance VIF

1. Gender .656 1.525 2. Age .024 42.200 3. Experience .085 11.699 4. Training .544 1.837 5. Overestimation .233 4.293 6. Flexibilty .013 75.148 7. Creativity .021 48.387 8. Environmental Awareness .029 34.902 Note: N=57

Table 4: Multicollinearity diagnostics of Model 2 with

competencies measured as difference from the mean

Variables Tolerance VIF

1. Gender .652 1.535 2. Age .060 16.733 3. Experience .090 11.147 4. Training .549 1.821 5. Overestimation .230 4.356 6. Flexibilty .601 1.663 7. Creativity .519 1.928 8. Environmental Awareness .568 1.762 Note: N=57

(28)

27

Regression Analysis

Model 1

As presented in Table 5, 6 and 7, a number of direct relationships were found between the three competencies and the variables. Training was positively related to creativity (β=0.18, P<0.01). Furthermore, the control variable age also significantly predicted the level of creativity, in which age was negatively related (β=-0.19, p<0.05). To examine the contribution of each predictor to the variance in the dependent variable (R2) results of relative weight analysis are also shown in Table 5,6 and 7. The Relative Weight Analysis (RWA) was performed using a macro which was developed by Johnson (2000). RWA measures the relative importance of each predictor’s ability to explain variance in the dependent variable in terms of percentage of R2. The results of the RWA show that the explained variance in creativity (R2= 0.053) can be attributed for 24% to Age and 54.3% to Training.

Furthermore, training also significantly predicted the level of environmental

awareness (β=0.21, P<0.01) and had a stronger effect than it did on creativity. The RWA also showed that a bigger part of the variance can be attributed to training than before (86.4%). While the percentage was higher, the total variance of environmental awareness that could be explained by the variables is lower (R2=0.043). No statistical significant relationship was found between training and flexibility (ß= 0.11, p>0.05) or any other control variables.

Table 5: Results of Training as predictor for Creativity.

Creativity Variables B SE B β Sig. % of R2 1. Constant 3.90 0.17 0.00 2. Gender _-0.15 _0.11 _-0.09 _0.15 _13.30 3. Age _-0.02 _0.01 _-0.19 _0.03 _24.00 4. Experience _0.01 _0.01 _0.15 _0.09 _8.30 5. Training _0.23 _0.07 _0.18 _0.00 _54.30 Note: N=284. R2=0.053.

(29)

28

These results support the hypotheses that propose that training improves creativity and environmental awareness, but opposes the hypothesis that training improves flexibility. What stands out from these three regression analyses is that the total variance that can be explained by the variables (R2) is considerably small. This strongly suggests that there are other predictors for the competencies, besides training in the case of creativity and environmental awareness.

Table 7: Results of Training as predictor for Flexibility.

Flexibility Variables B SE B β Sig. % of R2 1. Constant 4.19 0.13 0.00 2. Gender _0.02 _0.08 _0.01 _0.81 _0.53 3. Age _-0.01 _0.01 _-0.12 _0.18 _26.64 4. Experience _0.01 _0.01 _0.09 _0.30 _9.66 5. Training _0.11 _0.06 _0.11 _0.07 _63.18 Note: N=284. R2=0.016.

Table 8: Results of Training as predictor for Environmental Awareness.

Environmental Awareness Variables B SE B β Sig. % of R2 1. Constant 3.47 0.20 0.00 2. Gender _-0.03 _0.13 _-0.01 _0.84 _0.26 3. Age _-0.01 _0.01 _-0.10 _0.28 _11.10 4. Experience _0.00 _0.01 _0.05 _0.61 _2.20 5. Training _0.31 _0.09 _0.21 _0.00 _86.40 Note: N=284. R2=0.043.

(30)

29 Model 2

The relevant results of the regression analysis are presented in Table 8. Since the overall model with all the variables added did not prove to significantly predict the chances of successful succession (F = 1.352, p>0.05) backwards elimination was used to find a model that does significantly predict the chance of successful succession. To measure the contribution of each predictor, the Relative Weight is calculated by using a macro developed by Johnson (2000).

Initially, Table 8 shows that when all the variables are included in the model (Step 1) none of them appear to significantly predict the chances of successful succession, except the constant (B = 4.091, p<0.05). The variables which contribute the most to the total observed variance are: age (B = -0.001 ,p>0.05, RWA= 0.26), experience (B = -0.020 ,p>0.05, RWA= 0.327) and environmental awareness (B = 0.052 ,p>0.05,RWA= 0.157).

In step 3 of the backwards elimination process, only gender, experience, flexibility, environmental awareness and the constant remain. While age contributed 26 percent of the total predicted variance in step 1, the model apparently improved with its elimination. This could be explained by the presence of multicollinearity between age and experience, as was concluded from the correlation analysis. Experience proved to be a significant predictor of the chances of successful succession (B = -.020, p<0.05) and contributed the most to the variance (RWA= 0.531). Environmental awareness was not found to be significant predictor of the chances of a successful succession (B = .033, p=0.124>0.05) but did contribute 28.4% of the predicted criterion variance. This strongly indicates that environmental awareness could be a significant predictor of the chances of successful succession if a larger sample size was used, since it contributes a substantial part of the predicted criterion variance, is correlated with the chances of succession and is not significant by a minimal margin (0.124>0.05). Furthermore, flexibility was also not found to be a significant predictor (B= 0.127, p>0.05), but explained less than environmental awareness of the observed variance (0.111<0.294). The overall model in step 3 did not significantly predict the chances of successful succession (F = 2.087, p=>0.05).

Finally, in step 6, the overall model significantly predicted the chances of successful succession (F = 6.905, p=<0.05). Only experience (B = -.019, p<0.05) and the constant (B = 4.035, p<0.01) remain in this model. Apparently, experience has a weak but negative influence on the chances of a successful succession, but is nonetheless a significant

(31)

30

predictor. With only experience as a predictor variable, 11.2 percent of the observed variance is explained.

Table 8: Results of Flexibility, Creativity and Environmental Awareness as predictors for the Chances of Successful Succession using Backwards Elimination.

Variables B S.E. β Sig.

Relative Weight Step 1 1. Gender -.085 .194 -.064 .662 0.075 2. Age -.001 .014 -.009 .966 0.260 3. Experience -.020 .012 -.355 .101 0.327 4. Flexibility .131 .152 .135 .393 0.089 5. Creativity -.034 .129 -.047 .791 0.089 6. Environmental Awareness .052 .128 .075 .686 0.157 7. Constant 4.091 .367 .000 Step 3 1. Gender -.089 .182 -.067 .629 0.074 3. Experience -.020 .008 -.358 .010 0.531 5. Flexibility .127 .145 .131 .385 0.111 6. Environmental Awareness .033 .103 .047 .124 0.284 7. Constant 2.670 .806 14.442 .001 Step 6 3. Experience -.019 .007 -.334 .011 1.000 _{7. Constant} _4.035 _.144 _.001 Note: N=57. Step 1: R² = .140, F = 1.352, p=.252 Step 3: R² = .138, F = 2.087, p =.096 Step 6: R² = .112, F = 6.905, p =.011

(32)

31

Mediation analysis

Hypothesis 2a proposes that the relationship between training and the chances of successful succession would be mediated by the three competencies. Since the previous logistic regression analysis revealed that the competencies did not significantly predict the chances of a successful succession, it indicates that the competencies will not mediate the relationship. Still, the SPSS macro of Preacher & Hayes (2008) was used to examine the competencies as mediating variables. Results of the regression analysis are presented in Table 9, 10 and 11. First, training was not found to have a total significant effect on the chances of successful succession (β=0.568, p>0.05), nor did it had a direct effect. The small differences in the direct effect of training in Table 9, 10 and 11 are caused by differences in the used mediator and therefore the indirect effect. Since training did not have a significant total effect, it indicates that the competencies did not mediate the effect of training on the chances of successful succession.

This is supported by the indirect effect of training by flexibility, since the bias corrected confidence interval of the indirect effect of training on the chances of successful succession included zero (BC95 = [-0.176, 1.143]). Furthermore, the corrected confidence interval of the indirect effect of training by creativity also included zero (BC95 = [-0.212, 3.502]) and was not found to be significant (β = 0.005, p>0.05). This was also the case with the indirect effect of training by environmental awareness ( BC95 = [-0.108, 0.8388], β = 0.09, p>0.05) Thus hypothesis H2a, proposing a mediator effect of the three competencies on the relationship between training and the chances of a successful succession, is opposed.

Table 9: Regression results for Flexibility, as mediator of the relationship between training and the chances of a successful succession

95% CI

β SE β Sig. Lower Upper

Total effect of Training on the chances of Successful

Succession.

0.568 0.743 0.445 -0.889 2.024 Direct effect of Training on the

chances of Successful Succession.

0.461 0.765 0.547 -1.038 1.961 Indirect effect of Training on the

chances of Successful Succession. 0.102 0.310 0.594 -0.176 1.143 Experience -0.069 0.035 -0.052 -0.138 0.001 Note: N=57.

(33)

32

Table 10: Regression results for Creativity, as mediator of the relationship between training and the chances of a successful succession

95% CI

Succession.

Table 11: Regression results for Environmental Awareness, as mediator of the relationship between training and the chances of a successful succession

95% CI

Succession.