Founding-Family Ownership, Firm Performance and Crisis Resistance

Abstract

To further complete the currently fragmented findings of family firm research, this empirical thesis analyses the performance of large founding-family firms listed in the USA. Thus, this research builds up upon the studies of Anderson and Reeb (2003), Cannella et al. (2015), Miller et al. (2007; 2011), and Villalonga and Amit (2006). For the analysis, a random sample of S&P1500 firms in the period 1996 – 2018 was generated. To obtain information on family firm variables such as ownership and managerial involvement, a manual research was conducted for all 300 firms in the sample using annual reports and proxy statements. The impact of several aspects of founding-family involvement was analysed in a panel data analysis that employed the performance measures Tobin’s Q and Return on Assets as dependent variables. Controlling for industry and firm characteristics, the results did not back up the postulated hypotheses. The findings could not support an eventual outperformance by founding-family firms and challenged the results of past studies through contrary findings on the influence of firm founders and their descendants. The results of the analysis of Tobin’s Q and Return on Assets differ significantly and raise the question to which extent these differing results can be attributed to the stock market’s view on founding-family firms.

Introduction

Although there has been vast research on family firms and their performance over the past decades, no conclusive picture has emerged yet to answer the question whether family firms outperform their non-family counterparts and which factors are crucial for a superior performance. The results of previous studies vary substantially depending on factors such as the regional area of the analysis, the applied definition of the term “family firm” as well as on the analysed variables. Empirical findings on family firms in large public companies listed in the US demonstrate that family firms show a tendency to outperform their peers (e.g. Anderson and Reeb, 2003; 2004; Canella et al., 2015; McConaughy et al., 1998; Miller et al, 2007; Villalonga and Amit, 2006). In contrast to this, findings in other countries in Europe and Asia display results that speak for a lower average performance of family firms (e.g. Bennedsen et al., 2007; Claessens et al., 2002; Cronqvist and Nilsson, 2003; Maury, 2006). There are also discrepancies in performance when it comes to different sizes of family firms. Smaller family firms tend to underperform their peers (Holderness and Sheehan, 1988).

Due to these diverse, somewhat contradicting and unexpected findings, further research needs to be conducted and build upon the past findings to understand where the contradictory aspects are rooted in. Future empirical research must combine the existing findings and assess their overall validity to add to the whole picture of family firm research. Therefore, this empirical thesis builds up upon the studies of Anderson and Reeb (2003), Cannella et al. (2015), Miller et al (2007; 2011), and Villalonga and Amit (2006) and consequently focuses on large founding-family firms from the US and their performance. For this purpose, a panel data analysis of large U.S. companies was conducted by collecting data on a randomly selected sample of 300 S&P1500 firms in the timeframe 1996 – 2018. It was analysed if, how and for which constellations, a family influence has an impact on the performance of the firm. Performance was measured by both an industry-adjusted Return on Assets (ROA) and Tobin’s Q. Whereas the ROA was employed to assess performance from a current accounting perspective, Tobin’s Q rather represents the market’s evaluation of the firm’s prospects.

The thesis is organized as follows. Firstly, a short overview over the definitions of family firm in past research is given and the choice of the applied definition is explained. Secondly, the tested hypotheses are derived from a discussion of existing theory on family firms and past findings. Thirdly, the methodological part follows, in which the derivation of sample derivation and the involved variables are explained. The findings and main results of the analysis are displayed and explained in a fourth part ahead of a final discussion and conclusion.

What constitutes a founding-family firm?

In past research on family firms, various and sometimes widely differing definitions were applied for what constitutes a family firm. Definitions vary either on the level of involvement of the founders, their descendants and/or other family members. Additionally, different thresholds on the percentage of ownership of the family or individual members are applied. If the founder or the founder’s family is involved in the company by either serving on the board or in a management position or holding ownership of the company, the firm is frequently classified as a family firm in the literature (see for example Anderson and Reeb, 2003; Barth et al., 2005; Cannella et al., 2015; Cronqvist and Nilsson, 2003; Faccio and Lang, 2002; La Porta et al., 1999; Miller et al., 2007; Smith and Amoako-Adu, 1999) .

Other definitions require a stronger managerial involvement of family members, for instance one family member holding the office of the CEO (McConaughy et al., 1998). An involvement of the founder itself is also often treated as a special case and analysed separately. Miller et al. (2007) take this view on the importance of founder involvement one step further and distinguish between “lone founder” companies, in which only the founder and no other family member is present, and family firms, in which also other family members are involved. Their reasoning is based on the influences and possible conflicts of interests that result from the different responsibilities that a founder bears in his roles for the firm and for the family.

To allow for a comparison of findings, the definition of family firms in this thesis is oriented on the definitions of Anderson and Reeb (2003), Cannella et al. (2015), Miller et al. (2007; 2011) and Villalonga and Amit (2006). Firms in which all family members combined hold at least 5% of the voting rights or in which at least one family member is present as executive officer or director are considered to be a (founding-)family firm. The term family involves the founder or team of founders itself, their descendants or other family members related by blood or marriage. In the course of this thesis, the term family firm will be distinguished for further analysis into “first generation family firms” (founder and family members are involved parallelly), “second generation family firms” (only family members other than the founder are involved) and “lone founder firms” (only the founder is involved and no other family members).

Theory and hypotheses

Scholars in favour of an outperformance by family firms often lead the discussion by arguing with a mitigation of agency conflicts through the combination of ownership and control (Demsetz and Lehn, 1985; Fama and Jensen, 1983a; 1983b). This line of reasoning usually goes together with a presumably increased incentive of family firms for long-term value maximization (Anderson and Reeb, 2003; Casson, 1999) and correspondingly a longer investment horizon which increases investment efficiency (James, 1999). It also reduces the risk of myopic investment decisions which are often undertaken by external managers to increase short-term performance (Stein, 1988; 1989). Additionally, the reputation of the family is tightly connected to its firm. Therefore, the family has increased incentive to preserve a high company reputation and not to engage in fraudulent or unethical activities (Anderson and Reeb, 2003). This factor as well as their long-term orientation (see also Miller and Le Breton-Miller, 2005) might lead to a reduction of debt financing costs for family firms (Anderson et al., 2004).

On the other hand, authors that argue against a superior performance of family firms suggest that family firms are heavily impaired by an exploitation of the firm for private rents and gains for the family (Fama and French, 1983a; Shleifer and Vishny, 1997). In firms with family involvement, there is an inherent risk that the family chooses to maximize their own utility rather than the value of the firm itself (Bertrand and Schoar, 2006). This behaviour decreases the firm performance in the present as well as in the future as resources for growth are drawn from it (Demsetz, 1983). Correspondingly, executive entrenchment, the abuse of managerial power for own interest, appears to be very present in family firms (Allen and Panian, 1982; DeAngelo, H. and DeAngelo, L., 2000; Gomez-Mejia et al., 2001). Furthermore, findings show that family firms tend to show a weaker performance in certain aspects that are considered to be highly influential on overall firm performance. For example, family firms tend to display a lower productivity compared to their non-family counterparts (Barth et al., 2005) and focus less on firm value maximization (Bertrand and Schoar, 2006).

In the course of family research, scholars have also given increased attention to the role of the founder in family firms and their performance. In fact, the observed overperformance of family firms cannot be attributed to their status of family firm itself. Results of studies, in which it was distinguished whether the founder was actively involved, showed that an observed outperformance of family firms could not be connected solely to the firm’s status as family firm but rather to the founder’s involvement in the firm (Anderson and Reeb, 2003; Villalonga and Amit, 2006). Furthermore, it was analysed which position the founder or the family holds in the company and how this might impact performance. The results of these past studies show that if the founder holds the office of CEO, Chairman or the dual position of and CEO and Chairman (CEOCH), this tends to influence firm performance in a positive way.

In other, more recent research, the definition of the lone founder firm was introduced and analysed separately. A lone founder company is specified by the sole presence of the founder in the company whereas other family member have no stakes or influence in the company (Miller et al., 2007; 2011; Le Breton-Miller et al., 2011; Cannella et al., 2015). The reasoning behind this approach is based on a social point of view in which the lone founder – in contrast to the family firm founder – presumably focuses more on the interests of the firm and is influenced less by his role as a family member and the inherent social influence (Miller et al., 2011). Considering this distinction, lone-founder firms tend to outperform both “regular” family firms and non-family firms. The exclusion of the over-performing lone-founder firms from the group of family firms also leads to a relative underperformance of the latter group according to findings of Cannella et al. (2015) and Miller et al. (2007). Miller et al. (2011) had similar findings from which they concluded that especially lone founder owners and CEOs show a higher entrepreneurial orientation which consequently leads to an increased performance. In contrast, non-founder family owners or CEOs rather play a role as nourishers of the family and exhibit less entrepreneurial behaviour or traits which is reflected by lower firm performance. Family firms with their founders involved demonstrate a level of performance that lies between the one of the aforementioned types of firms. This type of founders seems to integrate both a familial and an entrepreneurial role.

Building on this discussion of past theories and findings in research, the following hypotheses were derived to be analysed via empirical models.

Hypothesis 1: Family firms demonstrate a better performance than non-family firms

Drawing on past findings (e.g. Anderson and Reeb, 2003; McConaughy et al., 1998; Villalonga and Amit, 2006), family firms in the U.S. display on average a better performance than non-family firms. Therefore, in this sample of S&P1500 firms, it is expected that there is a significant and positive difference in the analysis of performance of the family firms compared to the performance of non-family firms. Under the assumption of efficient markets for large publicly traded US companies, both Tobin’s Q and ROA are anticipated to display similar results in the analysis of the sample. This notion is supported by past findings which indicated superior performance of family firms for both the market-based performance measure Tobin’s Q (e.g. Villalonga and Amit, 2006) and the accounting-based measure ROA (e.g. Anderson and Reeb, 2003).

Hypothesis 2: Family firms outperform non-family firms in times of crisis

As elaborated, family firms tend to have longer investments horizons and are determined to secure the long-term survival of the company in the family’s own interest. Presumably, less agency conflicts among family firms reduce the risk of myopic investment decisions. Additionally, family firms tend to take risks to a smaller extent than non-family firms (Naldi et al., 2007). Due to their long-term perspective and relative risk aversion, family firms are expected to demonstrate increased stability and thus perform better in times of crisis compared to non-family firms.

Hypothesis 3: First generation family firms outperform second generation family firms and non-family firms

First generation family firms, in which the founder is involved in the company either by holding a position in management or on the board position and/or maintaining ownership over at least five percent of voting rights, perform better than non-family firms and family firms without founder involvement. This hypothesis is based on earlier findings and theories on the positive impact of founder involvement (e.g. Anderson and Reeb, 2003). Therefore, it is expected that there is a significant and positive difference in the analysis of the performance of first generation family firms compared to the one of non-family firms and second generation family firms.

Hypothesis 4: Lone- founder firms outperform all other categories of family and non-family firms

Drawing on past research on lone founder firms, this type of firm is expected to demonstrate on average a significantly improved performance when compared to other family and non-family firms. As discussed, this performance improvement has been attributed to the entrepreneurial role of the lone founder which distinguishes itself from the more familial roles in family firms and mitigates possible agency conflicts in non-family firms (Cannella et al., 2015; Miller et al., 2007; 2011; Miller and Le Breton-Miller, 2011).

Hypothesis 5: Firm performance is impacted positively if the founder maintains a strong managerial influence by holding the office of the CEO, of the Chairman (CH) or the dual position of CEO and Chairman (CEOCH)

Family firms in which the founder holds the position of CEO display superior performance compared to other family and non-family firms (Anderson and Reeb, 2003; Miller et al., 2007; 2011). Furthermore, it is expected that if the founder combines the two most influential positions in a company, a family firm not only outperforms non-family firms but also other types of family firms in which the founder is either not present or less influential (Miller et al., 2007).

Methodology and data

Sample Selection

The sample was built upon a random selection of 300 companies that were present in the S&P 1500 index at least in one year during the period of analysis 1996 – 2018. Data was exclusively collected for those periods in which each company remained a constituent in the S&P 1500 index. As the calculation of Tobin’s Q yields certain difficulties for financial firms, these companies, which are classified within the SIC code range 6000-6999, were excluded (Anderson and Reeb, 2003). This leads to a total of 2,371 firm year observations from which in 735 observations the company was classified to be a family firm equalling to 31% of total observations. In terms of firms, 234 companies were finally included in the analysis. Out of these, 91 were classified as family firms which represents 38,89%. The number of observations and firms varies slightly depending on the employed model due to missing data entries. Table 1 shows the detailed distribution of firms according to their 2-digit industry classification. Overall, the percentage of family firms and family firm observations appears in line with past findings in research on large US corporations (Anderson and Reeb, 2003; Miller et al., 2007, 2011; Villalonga and Amit, 2006). However, some differences in the distribution within the subcategories exist such as a high proportion of lone founder firms (100%) in “Industrial & Commercial Machinery & Computer Equipment”. An eventual attribution of these differences to the research period conditions will be analysed further in the discussion section.

Abbildung in dieser Leseprobe nicht enthalten

[...]