Introduction

In Latin America it is quite common that companies’ ownership is concentrated in family hands and business groups. As external corporate governance mechanisms - labor markets, legal framework, and markets for corporate control - do not function as properly as in developed economies (^{Boubakri et al., 2005}; ^{Chong and Lopez-de-Silanes, 2007}), firms rely mostly on internal corporate governance schemes- such as ownership concentration- in order to protect stakeholders and enhance firm value (^{La Porta et al., 1999}; ^{Steier, 2009}; Santiago, 2009). However, after the most recent financial crisis, more debate has arisen regarding the convenience of this entrepreneurial environment for firms´ performance, as it might limit access to capital markets and consequently increase companies´ risk levels (^{Barca, Ferri and Pesaresi, 1998}). Being Mexico the most injured Latin American economy during this crisis episode (according to ^{Bank of Mexico (2009)}, its GDP declined 6.5% during 2009), makes Mexican companies a provocative case to study. According to the latest numbers, most corporates listed in the Mexican Stock Market are family owned (^{Watkins et al., 2016}).

There is no unanimity in the literature regarding the business opportuneness of family control and management, particularly differentiating between normal times and crisis periods (^{Miralles-Marcelo et al., 2014}). Even though there are quite a few papers on the topic for developed markets, studies for Latin American economies are rare (^{Martínez et al., 2007}). This has to do, among other things, to the fact that detailed corporate governance information for companies operating in these countries is scarce. This paper adds to the limited literature on the Latin American context with the Mexican case, where typically there is high ownership concentration in family hands, incipient external corporate governance schemes, and therefore strong reliance on internal governance mechanisms in order to protect stakeholders and look after profit maximizing strategies.

Corporate governance and financial data for 89 non-financial companies listed in the Mexican Stock Market during 2001-2015 (which includes the crisis period between 2008 and 2009) has been captured manually, from the firms´ annual reports published at the Mexican Stock Market´s webpage. Controllers are found for all cases through company and internet search. Through this novel database, the relationship between ownership concentration in family hands and firm performance during normal periods and times of crisis is studied. Results for non-family owned versus family owned companies are obtained and compared, considering both the percentage of shares owned by the family (control) and its participation in the firms´ operations (management).

Results show that there are significant performance differences between family firms and non-family owned companies, considering return on assets (ROA) as the performance variable. Family firms outperform non-family owned businesses, while higher performance is appreciated at family corporates directed by family members rather than by outside CEOs. The latter effect is particularly strong during the crisis episode (2008-2009). Firm size, Board independency, and company´s age have a negative effect on return on assets, while ownership concentration is positively related with performance, especially during crisis times.

The paper is organized in the following way: Section 2 deals with the literature review and hypotheses. Section 3 describes the data, variables, and methods; Section 4 presents the main results, based on GMM estimations; Section 5 deals with additional endogeneity concerns; Section 6 concludes.

Literature review

Authors such as ^{Barca, Ferri and Pesaresi (1998)} and ^{La Porta et al. (1999)} put forward that high ownership concentration (for instance in family hands) restricts capital markets´ development, which limits access to financial resources and investment. The latter implies greater costs of capital, more entrepreneurial risk, and lower corporate performance. In addition, there is evidence in the literature that suggests that family owned companies´ outcomes are comparatively lower, relative to other proprietorship structures. In particular, when these businesses are managed by family CEOs, they tend to offer and sustain executive positions regardless of their descendants´ talents and administrative capabilities, which is not a good corporate practice (^{Galve, 2002}).

The negative relationship between ownership concentration (particularly in family firms) and companies´ outcomes is studied also by ^{Bennedsen and Wolfenzon (2000}); ^{Baek et al. (2004}); and ^{Maury (2006}). In accordance with the agency theory, it is shown in these studies that executives or majority shareholders expropriate minority shareholders and other stakeholders, with actions that evolve in lower earnings. The agency conflict arises as objectives are not always aligned. While owners search for profit-maximization, problems rise on issues such as the distribution of these benefits between minority and majority shareholders. As well, when majority shareholders are family entrepreneurs, they might search for socio-emotional objectives, which are not shared by other stakeholders (^{Gómez-Mejía et al., 2007}). The struggles between majority and minority shareholders are known as principal-principal problems (^{Jiang and Peng, 2011}). On the other hand, executives are also interested in their own personal economic benefit (^{Berle and Means 1991}; ^{Jensen and Meckling, 1976}), which creates divergences between their objectives and those of the shareholders (agent-principal problem). During times of crisis, due to the greater uncertainty and destruction of firm value, it becomes more attractive to expropriate different stakeholders (^{Mitton, 2002}); therefore, the negative effects on companies are greater. In emerging market economies where there is weak investor protection, lack of transparency and consequently more possibilities for expropriation, control becomes very valuable (^{Setia-Atmaja, 2009}; ^{Borhanuddin and Ching, 2011}). As stated by ^{La Porta et al. (2000}: 16): “In countries with poor protection, the insiders might treat outside investors well as long as future prospects are bright and they are interested in continued external financing. When future prospects deteriorate, however, the insiders step up expropriation, and the outside investors, whether shareholders or creditors, are unable to do anything about it”.

On the contrary, ^{Bunkanwanicha et al. (2008)}; ^{Boubakri et al. (2005)}; and ^{Kim (2006)} conclude that high ownership concentration (as in family hands) reduces agency costs and favors firm performance. Majority shareholders have greater incentives to monitor companies, in order to maximize profits; they might also hold an executive position which diminishes struggles between private and collective interests. In addition, in emerging markets families´ reputation might be needed in order to obtain external funding. High family ownership concentration favors long-term relationships in the companies, security and stability, together with knowledge transfer, which positively impacts investment and financial results - even during crisis times (^{Müller, 2014}; ^{Franks et al., 2008}, ²⁰¹²; ^{Hauswald et al., 2015}). Instead, in turbulent times, due to the circumstances, shareholders from widely dispersed-owned firms walk away from the corporations, which further deteriorates companies´ value (for instance, by selling shares with discount).

Business ethics and stakeholder relations are key elements in setting up good reputation and long term success of family firms (^{Dou et al., 2014}); therefore, it has been argued that even during crisis periods, expropriation of minority shareholders and creditors is not the rule but rather the exception (^{Bunkanwanicha et al., 2008}). A main goal of the family business is to ensure its own long-term survival, and therefore family entrepreneurs would be more interested than non-family ones in building up trust and taking into account the interests of all stakeholders (^{Le Breton-Miller and Miller, 2016}; ^{Miralles-Marcelo et al., 2014}). Due to long-term orientation, family-member CEOs would be expected to present less opportunistic behaviour, to make better use of resources and to seek out business strategies that maximise the firm’s value (^{Khanna and Palepu, 2000}; ^{Stanley and McDowell, 2014}). According to ^{Donaldson and Davis (1991)}, majority shareholders and family CEOs are well-intentioned individuals who want to avoid conflicts of interest and personal economic gains that can be harmful for other stakeholders.

A family-member CEO is expected to be more committed to the company, its present and future family generations, and to have more specific knowledge of the firm´s affairs. The greater alignment between the interests of family members and the family CEO could improve corporate performance (^{Jiang and Peng, 2011}; ^{Zellweger, 2007}). Furthermore, family-member CEOs usually have extensive family networks, connected to the worlds of politics and business, providing better access to information and resources, and reducing the uncertainty arising from emerging markets (^{Arregle et al., 2007}; ^{Sraer and Thesmar, 2007}).

In Latin America in general, and specifically in Mexico, family firms´ norms and reputation can substitute for the lack of stakeholders´ formal institutional protection through effective external corporate governance schemes. As reputation is a key element for companies´ long-term success, there are arguments to anticipate a positive relationship between family ownership, family management, and firm performance, bearing in mind both normal and crisis periods of time. As such, in line with the Latin American context, the following hypotheses are tested for non-financial companies listed in the Mexican Stock Market during 2001-2015 (including the 2008-2009 financial crisis period):

Hypothesis 1: Family firms outperform non-family owned businesses.

Hypothesis 2: There is a positive effect on performance when family firms are managed by CEOs who belong to the controlling families.

Data, variables, and methods

Methods

Given the characteristics of the database, it is possible to perform dynamic panel analysis. Panel data combines time series with cross sectional information. Panel analysis has the advantage that it allows to control for unobservable variables such as differences (that do not vary over time) in business practices between companies, as well as taking into account variables that change over time but not across firms (such as corporate law). As prior firm performance is an important determinant of current ROA (^{Tosi et al., 2000}), dynamics is attained by including in the regression lagged (one period) return on assets.

Regressions should also contain period effects (normal vs crisis), to contemplate performance differences between normal times and the most recent global financial crisis. For this purpose, a dummy variable (crisis) is introduced, being 1 if the years under consideration are 2008 and 2009, and 0 otherwise. Finally, in order to take into account the fact that ROA´s reactions due to financial and corporate governance arrangements might not be immediate, independent and control variables are lagged one period.

Corporate governance and finance literature argue the potential existence of endogeneity issues in the relationship between ownership and performance (^{Demsetz and Villalonga, 2001}), corporate governance (^{Pindado et al., 2011}), among others. In fact, family ownership might impact ROA, but ROA might also influence concentration of property in family hands. In order to deal with the endogeneity problem, regressions are run through the Generalized Method of Moments (GMM), for which a first differences transformation is employed in order to remove cross-section fixed effects. GMM level instruments are obtained then through habitual Arellano-Bond methodology. Robust, White period weights (2-steps) are used to compute standard errors. As stated by ^{Zhang (2013)}, GMM is useful to eliminate autocorrelation and heteroscedasticity problems, in addition to endogeneity (reverse causation) concerns.

The following equation is tested:

(1)

where:

δ₀ is the constant term;

δ₂ - δ₉ are parameters associated to each of the independent and control variables;

V refers to independent and control variables;

µ is the error term;

i refers to the companies;

t is time.

Equation 1is expanded to account for differences between periods. The dummy variable “crisis” (C) is introduced, being 0 for normal times and 1 for the crisis episode (2008-2009):

(2)

Results

Econometric results

Table 4 presents econometric results, for studying the impact of family control and family management on firm performance.

Table 4: GMM Estimations 

Source: Own elaboration

These outcomes are based on Panel Generalized Method of Moments (GMM estimations), using White period standard errors and covariance. As stated by ^{Zhang (2013)}, GMM is useful to eliminate autocorrelation and heteroscedasticity problems, in addition to endogeneity (reverse causation) concerns. As instruments, only stationary series are used, these being lagged first differences instead of variable levels. The latter follows ^{Kleibergen and Mavroeidis (2009)}.

Returnonassets(ROA) is thedependentvariable; independentvariablesarefamilyownership and CEO family. Control variables refer to ownership concentration, internationalization, firm size, leverage, Board independency, and company´s age.

The first two specifications include the whole time horizon, without differentiating between the crisis episode and normal times. In the first specification, equation 1 is run using as definition for family ownership when a single person or family owns 30% or more of the company´s ordinary shares. According to the results, family firms do not outperform (statistically speaking) non-family owned companies. There are no significant differences either when comparing firm performance according to CEOs´ family and non-family attributes. On the contrary, in specification 2, when shareholding rises to 51% or more in the definition of family ownership, a positive and highly significant effect of family firms on ROA is attained. In addition, there seems to be a slightly positive effect on ROA when family firms are managed by CEOs who belong to the controlling families.

With respect to control variables, both in specifications 1 and 2 there is an inverse size effect on return on assets. As the size of a business increases, its agency costs rise and, consequently, so do the possibilities of seeking objectives other than that of maximising business profit. Agents might be interested in investing in negative net present value projects, as firm size is related to compensation, power, reputation, and job stability (^{Jensen, 1986}). According to ^{Sharma et al. (1997)} and ^{Anderson et al. (2012)}, family owners would be interested in policies that favour socio-emotional wealth, even if they had a negative impact on the firm’s financial performance.

In specification 2, Board independency also appears to negatively impact firm performance. Although this is an uncommon outcome in the literature, it is not surprising as in Mexico the definition of independency does not take into account the fact that Board members tend to be related to other firms of the same business group (^{Watkins et al., 2017}). As stated by ^{Valenti and Horner (2010)}, this type of interlocking directorates allow a specific group of company owners, managers, chairs of the Boards, and other directors to unify corporate policy and concentrate power, which facilitates not the interests of the company itself, but rather the interests of particular directors. ^{Castañeda (2005)} shows that in Mexico the Board chair is usually the main shareholder and CEO, so this person in practice does not have opposition even from independent Board members.

For specification 2, it also shows that ownership concentration has a positive effect on ROA. High ownership concentration reduces agent-principal problems, as majority shareholders have the incentive to monitor companies in order to maximize profits (^{Khanna and Palepu, 2000}). In addition, large shareholders (family or not) have greater interest to have a firm running properly; business ethics and stakeholder relations are key elements in setting up good reputation and long term success. In these circumstances, the expropriation of minority shareholders as exposed in the principal-principal conflict is not the rule, but rather the exception (^{Bunkanwanicha et al., 2008}).

The third and fourth specifications differentiate between normal times and the most recent global financial crisis. For this purpose, a dummy variable (crisis - C) is introduced in equation 2, being 1 if the years under consideration are 2008 and 2009, and 0 otherwise. Specification 3 takes into consideration a family firm when a single person or family owns at least 30% of the shares; specification 4 considers a family firm when this shareholding is 51% or more. In both specifications, regression results manifest that family ownership positively impacts ROA during normal periods of time, while being the CEO a family member or not makes no significant difference. Regarding control variables, during normal times specification 4 displays that firm size, Board independency, and company´s age have a negative effect on performance. Mature companies tend to be less profitable, due to business cycle considerations (^{Ward, 2016}).

The effects of independent and control variables on firm performance during the crisis episode (2008-2009), corresponding to specifications 3 and 4, were corroborated through Wald tests. Table 5 indicates that during 2008 and 2009 there was no direct association between family ownership and ROA. In fact, the test values are negative (not significant), which moderates the positive overall impact of family proprietorship on firm performance when considering the whole time horizon in specifications 1 and 2. Nevertheless, during the crisis episode ROA was greater when the family firm was directed by a family member CEO, particularly for the stricter definition of family business used (51% of shareholding). In specification 4, three of the control variables appeared to impact return on assets during 2008 and 2009: ownership concentration being directly related with ROA; firm size and company´s age having a negative influence on performance.

Table 5 Wald Tests 

*** The default Wald test is rejected at 1% significance level.

** The default Wald test is rejected at 5% significance level.

* The default Wald test is rejected at 10% significance level.

1/ Family Ownership: A single person or family owns 30% or more of ordinary shares.

2/ Family Ownership: A single person or family owns 51% or more of ordinary shares.

Source: Own elaboration.

Endogeneity concerns

Endogeneity basically refers to the difficulty to determine a cause and an effect. The inquiry is if firm performance is the cause and family ownership the effect, or vice versa. We are interested in family ownership being the cause of higher Mexican firm performance. However, there is some evidence (see ^{Villalonga and Amit, 2006}; ^{Maury, 2006}) that families keep control only of well-performing firms (being performance the cause and family ownership concentration the effect). In addition, corporate performance can impact ownership concentration through mechanisms such as acquisitions taking place in the market for corporate control. The main concern with endogeneity is that one might under or overestimate the true effect of ownership concentration in family hands on firms´ performance.

Although endogeneity has been taken into account using GMM estimation, it is important to go further and use another approach to study the potential magnitude of this problem. For this purpose, it is analyzed if control changes are a common practice for quoted firms in Mexico. Since these types of events typically take place after performance variations (the reverse causality issue), they are related many times to CEO turnovers (see ^{Conyon and Florou, 2002}; ^{Fee and Hadlock, 2004}). In the same direction, when a company is acquired by a different controller, very often a new CEO is hired. Table 6 shows the complete list of CEO turnovers for companies quoted in the Mexican Stock Market during 2002-2015. From the table it is possible to conclude that in Mexico control changes are quite rare. In fact, from a total of 110 CEO turnovers, only 6 (5%) coincided with a change in ownership: BEVIDES 2002 (who was experiencing negative ROA from 2001 till 2002, and from 2003 on this was no longer the case), GMARTI 2007 (ROA was positive during all quoted years; the firms´ main shareholder and the Board of Directors accepted a tender offer), HOGAR 2007 (ROA was negative from 2005 till 2011), TMM 2009 (afterwards there was a positive performance effect), GMODELO 2013 (it was sold to Anheuser-Busch International Holdings), and AXTEL 2015 (it was suffering loses in 2014 and merged with ALESTRA). It is interesting to notice that there has never been a hostile takeover in the Mexican Stock Market, in part due to the high levels of family ownership concentration.

The information provided in table 6 suggests that in Mexico firm performance is not the cause for control changes, as property is quite stable; therefore, ownership concentration (family proprietorship) is more likely to impact company performance than vice versa. Hence, no further endogeneity concerns proceed.

Discussion and conclusions

This study evidences high family ownership concentration for quoted Mexican non-financial firms, which is typical in Latin America. Considering a family firm when a single person or family owns 30% or more of ordinary shares, on average 77% of the enterprises are family controlled. With a stricter definition, when a single person or family owns 51% or more of ordinary shares, this percentage still remains high (58%). In addition, 45% of these companies are directed by family members. These entrepreneurial characteristics do not vary significantly considering crisis and normal periods of time; in fact, they are quite stable variables.

Overall, family controlled firms outperform non-family companies. This intimate entrepreneurial environment benefits corporate results, as it promotes long-term relationships and responsible management (^{Henssen et al., 2014}; ^{Fernando et al., 2014}). In addition, CEOs who are members of the controlling families have a significant favorable impact on firm performance, particularly during crisis times. Family CEOs can reduce conflicts between managers and shareholders (^{Jensen and Meckling, 1976}; ^{Peng and Jiang, 2010}), which is beneficial for the firms. In addition, a family-member CEO is more committed to the company, and usually has more experience and information on the business (^{Bertrand and Schoar, 2006}). Also, family CEOs can access resources and reduce information constraints through their networks, which are above all beneficial when dealing with turbulent periods of time (^{Sitthipongpanich and Polsiri, 2015}).

In Latin American and other emerging markets, where external Corporate Governance mechanisms - such as the implementation of laws and regulations - do not function as efficiently as in developed economies (^{Boubakri et al., 2005}; ^{Chong and Lopez-de-Silanes, 2007}), firms rely more on internal governance schemes - such as family proprietorship - for stakeholder protection and consequently investor confidence (^{La Porta et al., 1999}; ^{Steier, 2009}). Families´ reputation is a key element for companies´ long-term success, which provides a counterbalance against expropriation of minority shareholders and creditors (^{Bunkanwanicha et al., 2008}; ^{Estrin and Prevezer, 2011}).

The combination of high ownership concentration in family hands, together with a family member CEO, is a strategy that excels financial outcomes in Mexican listed corporates. When the family owns an important percentage of the shares and participates in the administration of the company, it is easier to align objectives and reduce opportunistic behaviors, resulting in efficient resource and risk management (^{Galve, 2002}). This is especially relevant during crisis times, when asymmetric information and business risk increases.

The previous results suggest that Mexican family entrepreneurs act as responsible owners, taking into account the interests of present and future generations, as well as those of other stakeholders. Long time orientation - which is a characteristic of family firms - favors solidary leadership, stakeholder protection, and trust, which enhances firm performance (^{Brigham et al., 2013}; ^{Sharma et al,. 2014}). As an area of opportunity, future research on Mexican family firms should explicitly integrate time orientation-related variables.