The influence of the CEO in listed family businesses

Purpose: Our objective is to analyze the influence that the type of CEO has on the management of listed family businesses in Spain, distinguishing between whether the CEO is a family member or not. The study mainly focuses on his/her influence on levels of profitability. Design/methodology: During de period from, 2012 to 2016, with data coming from Iberian Balance Sheet Analysis System (SABI) database. To analyze the effects of the CEOs on family businesses, we carried out two kinds of analyses. First, a univariate analysis that allowed us to identify differences regarding profitability, financial structure, growth, and dividend payout policies, and secondly, a linear regression model to see the influence—as well as the effect and significance—that variables, including the type CEO, had on profitability. Findings: Our results show the existence of a double effect on the profitability of family businesses of having an outside CEO. First, there is a statistically significant negative effect that is derived from the non-family CEOs’ increased propensity to take on debt, and secondly, there is a positive causal effect on businesses’ profitability that has to do with the different management styles that outside CEOs bring to the table, as they are more focused on profits. The results support the importance of having non-family CEOs in listed family businesses in Spain. Research limitations/implications: Our study focused on family businesses listed on the Spanish stock market, which means that the number of companies that were analyzed was reduced and the results cannot be extended to other kinds of businesses. However, this fact did enable us to get more high-quality data and focus on a specific field that was appropriate for considering the problem we proposed. Originality/value: While many studies have compared the performance of family businesses with that of non-family businesses, few have considered that family businesses are not homogeneous and that they have different management styles. And, These styles are determined by the type of CEO that is leading the company; this fact is analyzed empirically in this article.


Introduction
When considering the characteristics of family businesses or the manner in which they act in comparison to non-family businesses, there is one particular factor concerning governance that often goes unconsidered: whether the CEO is a family member or not. These cases could have differing effects on the performance of the business.
From a theoretical point of view, the impact that these different types of CEOs can have on the level of achievement of the business is unclear. On the one hand, as some have suggested, family CEOs could achieve more than other executives as they receive non-monetary compensation, in addition to the normal benefits that any CEO would receive (Kandel & Lazear, 1992;Davis, Schoorman & Donaldson, 1997). It has also been argued that they have specific knowledge about the business that would be tough for an outsider to get and that they generate higher levels of confidence among stakeholders (Donnelley, 1964). Family executives could also benefit from having a long-term focus that non-family CEOs lack (Cadbury, 2000). On the other hand, some authors have suggested that family CEOs could actually perform worse due to the tension that arises from balancing family-and business-related objectives (Levinson, 1971; Barnes & Hershon, 1976;Lansberg, 1983).
Thus, the CEO's relation to the family represents a factor that could lead to differences when it comes to managing a family business. So, when the CEO is a family member, it is expected that he/she will make decisions focused on ensuring the future survival of the business and ensuring that the family maintains control of ownership. It is expected that an outside CEO will be more worried about producing positive results that attest to his/her abilities and will be less concerned with family interests. A non-family CEO will put profitability first, instead of maintaining so-called socioemotional wealth (SEW; Gallizo, Moreno & Sánchez, 2017a). The behavior of the outside CEO, however, could be subject to the control that owners exert over his/her managing of the company (Burkart, Gromb & Panunzi, 1997).
Our first hypothesis is that various priorities that drive the decision making of family CEOs and non-family CEOs lead to differences in management indicators in family businesses. Some of these differences should be able to be seen in policies related to financing, investment, and distributing profits. Furthermore, the managerial decisions made in one case or the other should have a direct effect on the performance of the business.
From our point of view, the greater profit orientation of an outside CEO, as well as his/her different management style in comparison to some family CEOs, lead to a situation in which an outside CEO running a family business will generate a greater return on assets (ROA). This is the second hypothesis that we test in the present article.
Considering our objectives, we used listed family businesses (on the Spanish stock market) from 2012 to 2016 as the sample for our study. We decided to focus on listed companies as they are required to provide more information and thus the classification between family and non-family businesses and CEOs was more reliable.
To test our hypotheses, we carried out two different analyses: first, a univariate analysis in which variables related to the businesses' financial structures, growth policies, and dividend policies were analyzed; and secondly, in order to determine whether the type of CEO had a significant influence on ROA, a linear regression model in which the type of CEO was included along with other control variables.
The results we obtained show the existence of differences in business management depending on whether the CEO is a family member or not. Regarding the influence of outside CEOs on ROA, we observed a two-sided effect: one negative due to their increased propensity to take on debt, and one positive due to their different management style and their focus on profits.
The rest of the article is structured as follows. In the next section, we discuss the differences that may exist between family CEOs and non-family CEOs as well as the motives behind their presence at the company. In the third section, we describe the sample and the models used in this study. In the fourth section, we present the results that we obtained, and in the last section we make some comments to summarize the main conclusions that can be drawn.

Family CEOs and non-family CEOs in family businesses
Although many studies have considered the differences that exist between how family and non-family businesses are run, not many have considered that family businesses do not represent a homogeneous business category. As suggested by Lin and Hu (2007), classifying all family businesses into one category could lead to biased conclusions, and thus other, operations-level characteristics (e.g., the type of CEO managing the company) must be taken into account.
In this way, and in the framework of agency theory (Jensen & Meckling;, the presence of an outside CEO could lead to agency problems and their corresponding costs. The principal-agent relationship involves the delegation of decision-making powers by the owners (principal) to the CEO (agent). If both parties are looking to maximize their own benefit, we can see that this situation would marginalize the CEO when it comes to making certain discretionary decisions that are not actually in line with the objectives of the owners. This could lead to the appearance of known agency costs between the agent and the principal: different motivation between them; asymmetric information, as the agent has specific, defined information about his/her work that the principal does not know; and predisposition to taking different levels of risk. And these costs are even greater in companies with dispersed ownership structures (Jensen & Meckling, 1976) than in companies with more concentrated ownership structures, as the principal can exert a greater level of control over the agent. Studies, such as those by Jensen and Murphy (1990) and Devers, Cannella, Reilly and Yoder (2007), have suggested that businesses should take actions to try to mitigate agency problems and their costs by implementing a remuneration scheme that is more apt for outside CEOs in such a way that the business' interests are more in line with those of the CEO.
Other works, such as that of Finkelstein and Hambrick (1996), maintain that the nature of the CEO could imply different managerial abilities, in addition to creating different agency problems (Jensen & Meckling, 1976). And thus, the choice of a family business to take on a non-family CEO depends on the characteristics of the organization and the managerial abilities required. Along this line, Lin and Hu (2007) suggest that when businesses require high levels of managerial skill, the performance of the business will improve if the CEO is not a family member. This result could be explained in part by the greater ease of finding qualified professionals in the job market than from within the family. Similarly, some articles suggest that occasionally, in family businesses, there is a certain willingness of family members to take on managerial roles even though they do not have the ability required to do so (Duréndez & García, 2005).
Thus, greater demands for executive abilities could require a separation between the ownership and the management, and this could be done via the hiring of an outside CEO (Burkart, Panunzi & Shleifer, 2003). However, the level of control that the family exercises over this CEO could have a negative effect on his/her performance, as any attempt to improve the business could be impaired by the family (Burkart et al., 1997).
Aside from the managerial abilities required, there are other factors that could affect the choice to take on one type of CEO or the other. One of the main factors is derived from the succession process in family businesses. It is more and more common for there to be no successor or family member who is able, willing, or qualified enough to accept the position (Chua, Chrisman & Sharma, 2003). As a consequence of this, it tends to be the biggest oldest businesses that have more experience with non-family executives.

Sample selection
The sample that we used in this study was made up of non-finance companies listed on the Spanish stock market from 2012 to 2016 and present in the Iberian Balance Sheet Analysis System database (SABI, by its Spanish initials). The sample originally included 102 companies but was later reduced when we removed companies that did not have the data necessary for study during all of the years considered. We also removed all companies that showed negative equity capital in any year, as we considered these to be atypical situations that could distort the results of the study. Finally, we removed companies that, during the period considered, underwent ownership changes that affected their classification (family or non-family) as well as those that changed the classification of CEO (family CEO or non-family CEO). After applying these filters, we were left with a final sample of 78 businesses.
We obtained economic and financial data, as well as information related to businesses' ownership structures, from the SABI database. Regarding information on the CEO, we found some errors in the aforementioned source, so, we assessed the information with websites which provide this data with greater levels of veracity.
To classify businesses as "family" or "non-family" businesses-one of the main issues in this kind of study-a definition of "family business" had to be adopted. In Europe, there are more than 90 such definitions (Mandl, 2008). While they all address various different dimensions and might differ to some extent, they all focus upon three key aspects: ownership controlled by one family, family participation in the management of the business, and the willingness to pass on ownership to the next generation.
For this study, we decided to use the definition suggested by the Instituto de la Empresa Familiar (Family Business Institute, 2015). In their study, they considered the difference between businesses with a dispersed ownership structure versus those with a well-concentrated ownership structure to be the main factor in establishing an operational definition of a family business. These differences in ownership structures have led to a debate over what ownership percentage is appropriate for classifying a business as a family business. Along this line, we felt that it was not appropriate to use the same percentages for all businesses, as in those with a more disperse ownership structure; a high ownership percentage is not needed to exert control over the company. As such, we used the criteria adopted by the Family Business Institute and the information available in the SABI database to set up the following definitions: • Disperse ownership structure (no shareholder has more than 50% of capital). The family business will be that in which one person owns 5% or one family owns 20% of the capital, and the natural person shareholder is a board member or the family represents shareholders with more than 20% of capital and governing power. Otherwise, the business will be classified as a non-family business.
• Concentrated ownership structure (some shareholder has more than 50% of capital). The family business will be that in which the family-member shareholder controls a high percentage (50.01%) of ownership, or there are shareholder-board members with more than 50.01% involvement. Businesses not fulfilling these conditions will be classified non-family businesses.
Once we applied these criteria, we conducted a revision in order to classify those businesses for which ownership information was incomplete and/or unclear, and to revise the assignments that had been made.
The classification of the businesses in our sample can be seen in Table 1. The majority were family businessesrepresenting 61.54% of all of the businesses considered. This proportion is in line with the data that has been published by the Family Business Institute. Family businesses represent around 90% of all businesses in Spain, but this proportion decreases significantly among larger companies, which is obviously the case for the listed companies that were the object of this study. As for the type of CEO in family businesses, the results showed similar proportions of family and non-family CEOs, with a slight majority being from outside of the family. This result matches with the most recent data published by the Family Business Institute (2018): while the large majority of CEOs in family businesses are indeed family members of the owning family, the rate of family CEOs decreases as the size of the company increases, becoming almost even among large-sized businesses. This characteristic is not unique to Spanish companies. In the USA for example, 20% of family businesses have an outside CEO (MassMutual American Family Business report, 2003), and this number increases to 55% in the 141 large family businesses listed on the S&P 500 (Anderson & Reeb, 2003). Similarly, in Germany, the percentage of non-family executives on management teams increases along with the size of the family business (Klein, 2000).
Thus, our study focused on the 48 family businesses for which it was possible to define the type of CEO (family or non-family) and the influence that he/she had on the performance (especially ROA) of the business.

Definition of the model
To analyze the influence of having a family or outside CEO on the profitability of family businesses, we estimated a multiple linear regression model. The variables we used are described hereafter and are summarized in Table 2.

• Dependent Variable
As an indicator of the profitability of the company, we used return on assets (ROA), calculated as income, before expenses and taxes, over total assets. This variable does not include capital costs, instead reflecting the results achieved by the company via their use of assets. Many studies have used ROA as a variable that is representative of the performance of a company (Stickney, Brown & Wahlen, 2007), also for family businesses (Anderson & Reeb, 2003;Andrés, 2014;Arosa, Iturralde & Maseda, 2010;Lam & Lee, 2008).

• Independent Variable
As the independent variable, and the object of our study, we included the "type of CEO." This took the form of a dichotomous variable that had a value of 0 if the CEO was a family member and a value of 1 if the CEO was not a family member. The type of CEO could entail different managerial skills (Finkelstein & Hambrick, 1996) or various kinds of agency problems (Jensen & Meckling, 1976), which would undoubtedly affect company performance. The ability of CEOs is vital to the performance of businesses when they require high levels of managerial skills (Burkart et al., 2003). And thus, in the listed companies analyzed in this study, we expected that an outside (non-family) CEO would contribute to attaining better outcomes.

• Control Variables
• Sector of activity: All companies that have been accepted for listing on the Spanish stock market and can be traded via the Spanish Stock Market Interconnection System (SIBE, by its Spanish initials) or by the Corros system, are classified into one unified sectoral and sub-sectoral framework that was introduced on 1 January 2005. This framework defines six basic sectors: petroleum and energy; basic materials, industry, and construction; consumer goods; consumer services; financial services and real estate; and technology and telecommunications. Given that we excluded companies from the finance sector from the study and that there were no family real estate businesses in the sample, the model allowed for the existence of the five remaining sectors. As a consequence, five dummy variables were included in the model, with the basic materials, industry, and construction sector (the largest sector) serving as a reference. Then, the "petroleum and energy" variable took the value of 1 if the business belonged to this sector and 0, otherwise. The "consumer goods" variable took the value of 1 if the business belonged to this sector and 0, otherwise, and so on successively with the rest of the sectoral variables. As a result, the regression coefficients for these variables represent the differential effect of each sector on the dependent variable (ROA) in relation to the reference category (basic materials, industry, and construction). Previous studies have confirmed the existence of differences depending on the sector in which the business operates (Stickney et al., 2007).
• Year: Five dummy variables corresponding to the respective years under study (2012, 2013, 2014, 2015, and 2016) were included, with 2012 being used as the reference year. In this way, the "2012" variable took the value of 1 if the observation corresponded to this year and 0, otherwise. The "2013" variable took the value of 1 if the observation corresponded to this year and 0, otherwise, and so on successively for the rest of the years considered. As a result, the regression coefficients for these variables represent the differential effect of each year on the dependent variable, with regard to the reference category (2012). As pointed out by Salas-Fumás (2014), the levels of ROA of non-finance companies in Spain decreased during the first years of the economic recession (2007)(2008) and progressively recovered and reached their highest levels in the following years. Nevertheless, given that the period of time being considered in our study could still include effects from this economic recession, we felt it was necessary to control for time in our analysis.
• Total assets (Ln(Assets)): Calculated as the natural logarithm of a company's total assets so as to minimize asymmetry of the variable given its high level of variability, its size is related to many of a business' characteristics, and for this reason, it is often included as a control variable (Anderson & Reeb, 2003;Andrés, 2014;Carter, Simkins & Simpson, 2003;Barontini & Caprio, 2006;Arosa et al., 2010). Previous studies have found a negative relationship between the size of the company and performance (Lang & Stulz, 1994).
• Age of the company (Ln(Age)): Measured as the number of years since the founding of the company, we took the natural logarithm of age in order to minimize asymmetry, given the high level of variability found in this variable. The inclusion of age as a control variable is common in the literature (Andres, 2014;Arosa et al., 2010;Cabrera-Suárez & Martín-Santana, 2015). It is seen as a measure of a company's ability to compete in a highly competitive environment. Previously, Evans (2007) observed a positive relationship between age and profitability, though Cooley and Quadrini (2001) affirmed that the growth of a company decreases as age increases. Similarly, Shleifer and Vishny (1989) suggested that in family businesses, getting bogged down with family issues could lead to founders remaining active in the company, even though they are not competent enough to do so. This is common in business with concentrated ownership structures.
• Indebtedness (Debt): Measured as the total debt of the company over its total assets. Andrés (2014) and Arosa et al. (2010) identified a significant negative influence of the level of debt on the profitability of non-listed Spanish businesses. Thus, we expected to find a similar relationship in listed companies.
• Liquidity ratio (Liquid): Calculated as the current assets over current liabilities. This measure allows us to analyze the ability of a business to confront their debts in the short term. Its inclusion as a control variable is common (Gul & Leung, 2004;Lam & Lee, 2008).
• Growth of sales (GSales): Calculated as "year x sales / year x-1 sales." Previous studies, including Scherr and Hulburt (2001) and Arosa et al. (2010), have included this variable in their models as businesses that grew more in the past have greater possibilities of growing in the future, and thus, influencing ROA.

Univariate analysis
Before estimating the model that allowed us to identify the influence of the type of CEO on a family business' ROA, we analyzed a set of economic and financial variables in order to see whether or not there were managerial differences between the CEOs. Specifically, we took a look at the variables related to profitability, financial structure, growth, and dividends policy.

Return on assets
Regarding ROA, we observed that businesses with an outside CEO tended to perform better than businesses with a family CEO, though this trend was not found in all of the years considered (Table 3 and Figure 1).  The greater average profitability seen in family businesses run by non-family CEOs suggests that these executives have better managerial skills and a greater tendency to prioritize financial results over other kinds of objectives. Nevertheless, given the influence that various, potentially internal, factors could have on levels of profitability, a univariate analysis is not the most appropriate for corroborating these conclusions. It is for this reason that we decided to perform an econometric analysis that allowed us to isolate the affect that the type of CEO has on the profitability of these businesses. Table 4 shows the evolution of the average indebtedness of businesses by type of CEO (family or non-family). These results suggest that family CEOs prefer self-financing, a preference that is justified by their greater apparent risk aversion and prioritizing maintaining control over the company. This behavior by family CEOs is in line with behavior that is traditionally associated with family businesses, i.e., they prefer to finance with their own family money (Gallizo, Mar-Molinero, Moreno & Salvador, 2017b;Hamilton & Fox, 1998;Romano, Tanewski, & Smyrnios, 2000) or with undistributed profits (Hamilton & Fox, 1998;Barton & Matthews, 1989) instead of using debt or bringing on new shareholders. On the contrary, the results obtained for non-family CEOs show evidence of their preference for taking on debt, a practice that is more in line with non-family businesses. For a more in-depth analysis we calculated the bank debt ratio. This ratio takes into account the type of debt which generates financial expenses for the company, and consequently, the type of debt which directly affects the results of the company. The results are shown in Table 5 Table 5 corroborate what we previously saw regarding levels of debt: non-family CEOs opt for greater levels of bank debt ratio in all of the years considered in this study.

Financial structure
In order to analyze the financial structure, we also calculated the liquidity ratio, which shows the ability of a company to pay off debt obligations in the short term. Table 6 shows the results of this analysis. We can see that in all of the years analyzed, except for 2014, businesses with a family CEO had greater levels of liquidity compared to businesses with a non-family CEO. As before, these results show that family CEOs tend to have a greater risk aversion than non-family executives and they prefer financial autonomy even though this may occasionally mean lower levels of profitability as a result of not taking full advantage of available resources.

Growth
In this section, we take a look at the differences in growth that businesses saw depending on the type of CEO that they had. Specifically, we show the growth in total assets (Table 7) and the growth in noncurrent assets ( Table 8).
As we can see, businesses with a non-family CEO saw greater levels of asset growth throughout the period considered, and this growth was positive in all years. Contrarily, businesses with a family CEO saw these levels decrease, on average, over the years analyzed.
Characteristics that we saw in businesses with a family CEO mirror those that have been observed in previous studies on family businesses, where it is suggested that such businesses prefer control, which limits their possibilities for growth (Galve & Salas, 2011 To make a more in-depth analysis, we also considered the growth of noncurrent assets (Table 8). The results of this analysis once again show that businesses with non-family CEOs saw greater levels of growth over the period analyzed. These results are not surprising given that investments in noncurrent assets are generally financed with long-term funding. Previously, we showed that businesses with family CEOs have lower average levels of debt. These businesses also tend to limit the entrance of new associates in order to maintain control of the company. Both of these facts restrict their investment abilities (Galve & Salas, 2011).

Dividend payout
Finally, we calculated the average payout of businesses with family CEOs and those with non-family CEOs (Table 9). The results we obtained show that, over the period considered (with the exception of 2015), businesses with a non-family CEO distributed a larger portion of their profits as dividends.  Table 9. Payout (Dividends/Profit) for businesses with profit by type of CEO Previous studies have suggested that there is a hierarchy that exists in family businesses when it comes to their financing preferences: they are more averse to risk and implement more restrictive dividend policies (Gallizo et al., 2017a;Romano et al., 2000). Along this line, family CEOs tend to favor reinvesting profits in order to finance the growth of the business and ensure its survival. On the contrary, non-family CEOs tend to favor distributing dividends, which attests to their managerial abilities.

Regression results
First, we estimated a model without control variables related to the financial structure of the businesses. We did this because, as per the previous section, we saw that the financial structure of a business is largely determined by the type of CEO that is running it. In line with previous studies that have shown the influence that these variables have on ROA (Anderson & Reeb, 2003;Andres, 2014;Arosa et al., 2010;etc.), inclusion of this variable could mask part of the influence that the type of CEO has on profitability. Results corresponding to the estimation of this model are shown in Table 10.
With regard to the variable being studied-the type of CEO-we can see that its influence, though positive, is not statistically significant, which would lead us to draw the conclusion that the type of CEO (family or nonfamily) does not influence the levels of profitability of the businesses analyzed.
With regard to the control variables, the most noteworthy results are the strong influence that the sectors of activity have on ROA. With the basic materials, industry, and construction as a reference, we can see that the consumer goods sector has a positive and statistically significant influence on ROA (at the 1%-level). This indicates that businesses in this sector tend to see greater levels of profitability than businesses in the reference sector. On the other hand, companies in the petroleum and energy sector and those in the technology and telecommunications sector show a statistically significant (again, at the 1%-level) negative influence, suggesting that businesses that operate in these sectors have lower levels of profitability than those operating in the reference sector.
As for the year in which the observations were made, none of the dummy variables included in the model had a significant influence on ROA, suggesting that there were no significant differences between the years in the period analyzed, with regard to profitability. This result is in line with the work done b y Salas-Fumás (2014), where ROA in Spanish businesses was seen to have decreased at the beginning of the economic recession (2007)(2008) but recovered over the following years with no major differences since then.
Among the rest of the variables considered, only one was statistically significant (at the 5%-level): the age of the company. Specifically, we can see that this variable has a negative sign, suggesting that businesses that have existed for longer will see lower levels of ROA than younger businesses. This result has been documented previously by Andrés (2014).
None of the other variables (natural logarithm of total assets and sales growth) had a significant influence on ROA. Hereafter, we present the results for the estimation of the model with the variables related to the financial structure of the company included, specifically the debt ratio (liabilities/total assets) and the liquidity ratio (current assets/current liabilities). This allowed us to see whether there were changes in the influence that the type of CEO had on ROA when we isolate the effect that the variables related to the financial structure may have had. The results of this regression appear in Table 11.
As we can see below, the inclusion of these two new variables has slightly improved the goodness of fit of the model. Furthermore, we can see that the indebtedness has a significant negative influence (at the 1% level). So, this result confirms that there is a negative relationship between a company's debt and its profitability, a result that is in line with the findings of previous studies (Andres, 2014;Arossa et al., 2010). Table 11. Coefficients of the entire model Once we isolate the effect of the level of debt, we can see the influence that the "type of CEO" variable has is statistically significant (at the 10%-level). Thus, having a non-family CEO contributes to improving the ROA of a business. This result suggests that non-family CEOs have certain abilities that contribute to the improvement of profitability that were not seen in the previous model because of the negative influence that the level of debt (which was previously captured by the "type of CEO" variable) had in the regression.
The results that we obtained in these two models reveal a double effect of having a non-family CEO on ROA. First, there is a negative effect that comes from their preference to take on debt, which has been shown to worsen the profitability of companies. And then, a positive effect, derived from the non-family CEO's managerial abilities and focus on profits.

Discussion and conclusions
The present study analyzed the influence that the type of CEO (family or non-family) has on a set of variables related to the management of family businesses, and especially their influence on ROA.
To do this, we analyzed a sample of 48 family businesses that were listed on the Spanish stock market from 2012 to 2016. First, we carried out a univariate analysis of the set of economic and financial variables in order to see whether there were actually any differences in the way that family CEOs manage compared to non-family CEOs.
The results of this analysis revealed that there were specific differences in the management styles of these CEOs of family businesses, especially regarding policies on debt, growth, and dividend payouts. We found that the level of total debt was greater in family businesses with a non-family CEO, which suggests that these kinds of executives have a greater willingness to take on debt than family CEOs. This kind of performance is more in line with the behavior of non-family businesses. Similarly, businesses with a family CEO maintain greater levels of short-term liquidity, suggesting their greater aversion to risk and preference to have greater financial autonomy.
With regard to growth, the results show that family businesses with a non-family CEO obtain greater investment rates than family businesses with a family CEO. On average, businesses with a family CEO decreased in size due to growth restrictions stemming from their characteristics as a family business with a family CEO. Also, with respect to dividend payouts, our results show that businesses with a non-family CEO tend to distribute a greater proportion of their profits as dividends; this is in line with the hypothesis that family CEOs generally opt more for reinvesting profits in the company itself.
Finally, to analyze the influence of the type of CEO on ROA, we estimated a multiple linear regression model in which we first estimated the influence of the type of CEO without including control variables related to the financial structure of the business (variables that are closely related to the type of CEO) and then with these variables (specifically, the debt ratio and the liquidity ratio) included in order to see whether isolating their effect could change the observed influence of the type of CEO on ROA.
The results of the initial estimation showed the absence of a significant effect of the type of CEO on ROA, however, based on the results of the second estimation, we were able to deduce that the greater levels of debt that family businesses with a non-family CEO take on negatively influence their levels of profitability. Then, when this negative effect was isolated, we were able to see the statistically significant positive effect that a nonfamily CEO had on the businesses, which could be explained by their managerial skills and greater focus on profits. Ultimately, the results of our study speak to the existence of a double effect that non-family CEOs have on the profitability of family businesses.
The main shortcoming of this study lies in the reduced number of businesses available for consideration, as we focused exclusively on listed family businesses in Spain. Future research should therefore expand the scope of similar analyses to include a larger sample of businesses, possibly including non-listed companies. Another interesting question that should be considered is whether the president of the business takes on an executive role or not, as if he/she did, it could restrict the independent management of a non-family CEO.