پرداخت غرامت CEO و سهامداران بزرگ: شواهدی از بازارهای نوظهور

Using a novel data base for three emerging markets we study large shareholders and their relationship with professional managers. This is important to understand wage inequality and returns to high-level human capital since concentrated ownership is prevalent in developing countries. We find a compensation premium of about 30 log points for professional (not controller-related) CEOs working in firms controlled by a family compared to firms controlled by other large shareholders. The premium cannot be explained away by standard firm characteristics, observable executive skills (e.g., education or tenure), or the compensation of the CEO in her former job. The premium comes mostly from family firms with absent founders and when sons are involved.

There is a large literature on the effect of institutions – e.g., the legal system, the government, or the country’s cultural background – on economic performance. The institutions that prevail in the corporate sector also vary substantially across countries. Perhaps one of the main differences between the US corporate sector and other countries is that large controlling shareholders are rarely seen in the US while they are the norm in other countries. As suggested by Morck et al. (2005) the concentration of ownership in large shareholders can have macroeconomic impact, for example, by affecting resource allocation, public policies, and the distribution of income and political influence. The concentration of corporate ownership is typical of Latin America – our focus of study – but also Asia and continental Europe (see Claessens et al., 2000, Faccio and Lang, 2002 and La Porta et al., 1999).1 Beyond ownership concentration, Latin American financial markets are similar to many of these other markets in terms of their size (relative to the country’s GDP), the number of listed firms relative to population, the legal protection given to investors, and other dimensions (Djankov et al., 2008). Therefore, understanding corporate institutions in Latin America can shed some light on a host of other markets. In this paper we study large shareholders and their relationship with professional managers. Given the debate about the role of large shareholders in executive compensation, and in corporate governance more broadly (see, for example, Shleifer and Vishny, 1997), this paper attempts to fill a gap in the literature by studying whether it is appropriate to talk about large shareholders as a uniform class of “hands-on” principals or if different large shareholders have different implications for how managers are compensated.2 In particular, we study whether families stand out from other classes of large shareholders in terms of how they compensate executives. In addition, understanding the relationship between the ownership structure of firms and managerial compensation can help us to understand in a better way the returns to high-level human capital in developing countries. These countries typically present high wage premia. The principal-agent framework has been the workhorse model for studying executive compensation as seen in Jensen and Murphy, 1990 and Murphy, 1999. In this framework a principal must delegate control to an agent who cannot be easily monitored, and who needs incentives to work hard. This captures the conflict between a set of dispersed shareholders, with little incentive to monitor, and a manager, which is typical of US firms. Principals with large stakes, which are more typical of the corporate environment in Latin America, have powerful incentives to monitor managers and therefore can be a solution to the principal-agent problem posed by dispersed ownership (Shleifer and Vishny, 1986). For example, Bertrand and Mullainathan (2001) find that executive pay is less sensitive to pure luck in firms with large shareholders. A parallel literature studies the peculiarities of one form of concentrated ownership: family ownership (e.g., Bertrand and Schoar, 2006). Family firms are special in that they tend to pursue a special set of values, managerial practices, and specific traditions related to the family (sometimes with a strong distrust for outsiders). Management is many times kept within the family, even after the retirement of the family founder, potentially as a way to preserve a specific form of human capital related to the family’s business expertise. Family ownership may also act as a remedy for market imperfections that exacerbate agency problems (e.g., Burkart et al., 2003). All these dimensions make family ownership stand out from other forms of concentration and of particular interest in emerging markets (Fan et al., 2011). The lack of empirical work on this topic is probably related to the absence of detailed data on executive compensation outside the Anglo-Saxon world precisely where large shareholders are prevalent.3 As emphasized by Fan et al. (2011), due to the opacity of data we still do not know much about managerial compensation in emerging and developing markets. In this paper we shed light on this issue by presenting a unique data set of approximately 1700 top executives in Argentina, Brazil, and Chile. For each executive we know base and bonus compensation, biographical information such as age, gender, education, and a detailed description of previous work experience. For approximately 40% of executives we also have compensation in their previous jobs. Through the biographical information we can study several dimensions of executives’ careers such as tenure and promotion within firms, which complement the study of compensation itself. The executives in our sample work in a wide array of firms: private and publicly-traded, small and large, in financial services and manufacturing, and so on. Despite this variety, ownership structures are amazingly homogeneous. The majority of firms are controlled by a single, easily-identifiable, large shareholder, who typically is a family, a foreigner (mostly foreign corporations), or the government. The rest of the firms are controlled by coalitions of a few wealthy individuals, families or foreigners. Widely-held corporations, in which there is no controlling block of, say, 10% of shares or more, are almost non-existent. Some could argue that these markets are perhaps not the best setting to study executive compensation precisely because of the high levels of ownership concentration. However, if one wants to understand the role of large shareholders in executive compensation, then one could even argue that these are among the best markets to study their influence. Also, these Latin American countries are at the top of the pack of emerging markets, instead of being among the relatively poorer ones. This ensures that the companies that we study are modern corporations, comparable in many ways (not all) to corporations in developed countries in terms of organizational structure, internal hierarchy, managerial practices, and so on. Our main result comes from comparing executive compensation across different classes of large shareholders. We find that professional CEOs in family-controlled firms make around 30% more than CEOs in other firms. The premium is not observed among executives below the CEO level. Given that our sample includes only professional executives who are not family members, the family-premium is not a mechanical result of nepotism. We explore several hypotheses that could explain the family premium. First, the premium can reflect special characteristics of family firms, for example, that family control is more prevalent in certain sectors, such as financials, where executive compensation is typically higher. We do not find evidence for this. In fact, the family premium survives a host of control variables such as sector fixed effects, firm size, volatility, and profitability. The fact that the premium is seen only among CEOs and not among other executives also suggests that the premium is not a firm-level effect, but something more specific to the CEO position. Second, it can be the case that CEOs in family-controlled firms have special characteristics vis-à-vis executives working for other types of large shareholders. At least in terms of observable characteristics we do not find clear evidence for this either. CEOs in family firms are of about the same age, education, and are as likely to come from the lower ranks of the firm as CEOs in other firms. Executives in family firms frequently come from firms controlled by other types of large shareholders and viceversa, which reinforces the idea that executives in family and non-family firms are part of a general executive population and not separate populations. The model of family firms in Burkart et al. (2003) suggests that families hire high-quality CEOs since these make the delegation of management more attractive.4 The fact that we do not find a clear difference in observable executive skills, such as education or experience, seems to be against this hypothesis. However, CEOs in family firms can still be more talented than executives in other firms. For instance, they can have soft skills (unobservable to the econometrician) that justify the compensation premium. For example, one can imagine that CEOs in family firms need special skills to interact with the family members, who potentially have little formal business education. The greatest obstacle to test this hypothesis is to find an empirical measure for unobservable skills. We follow Gibbons and Waldman (1999) who argue that compensation in previous jobs can be a proxy for the sum of observable and unobservable skills.5 We find that the family premium is still of about the same magnitude after we control for former compensation, and that former compensation itself does not have a significant impact on current compensation among CEOs (it does have an impact in other executive levels). Similarly, higher current compensation does not predict being hired by a family in the future. Again, this evidence does not support the idea that the family premium is a compensation for managerial skills, although it leaves the door open for other soft skills that are not captured by former compensation. A third possible explanation is that family firms are different in the way the principal-agent relationship is dealt with. Hands-on principals need to offer higher direct compensation because of the detrimental impact they have on managerial private benefits. More intense monitoring curves perk-taking, but for the same reason it reduces the incentives of executives (Burkart et al., 1997). In order to retain incentives managers must receive higher direct compensation. The interpretation of our finding would be that families are particularly strict monitors. Hands-on principals can also have an impact on private benefits in more subtle ways, for example by affecting the executive’s own human capital and career. Rajan and Zingales (1998) argue that agents increase their (inalienable) human capital by receiving access to critical resources controlled by the principal. Empirically we find that the family premium comes mostly from firms where the founder is absent, which can be a sign that managers ask for a compensation if they do not have access to the business expertise of the founder (as suggested by Bertrand and Schoar (2006), having a business-savvy founder is arguably the critical resource behind the success of many family firms). Similarly, we find that the involvement of sons of the founder in management or the board of the company is associated with a larger premium in CEO compensation. The presence of sons can increase the chance of CEO replacement and damage the CEO’s career, although we do not find significant differences in the tenure of CEOs in family firms when compared to other firms. Still, the perception of higher career risk may lead CEOs to ask for higher compensation. Finally, the family premium can represent rent extraction on the part of CEOs (Bebchuk and Fried, 2004). Perhaps family firms are more easily captured by a professional CEO than other large shareholders. The fact that the family premium is seen mostly when the founder is absent seems to support this idea, because other family members may lack the experience of the founder. We do not find conclusive evidence in our sample to discriminate between these different mechanisms, and in reality they can all work simultaneously or apply to some firms more than others. However, an important take-away from our results is that the class of large shareholder matters for CEO compensation. In other words, not all forms of ownership concentration are the same when it comes to hire and compensate managers. In particular, if the large shareholder is a family then CEO compensation is higher on average than in a comparable firm controlled by a different class of large shareholder. This result is related to the finding of Cronqvist and Fahlenbrach (2009), who report that large-shareholder fixed effects are significant among US firms for explaining executive compensation among other corporate policies. We go one step further in predicting that families are the large shareholders with big and positive fixed effects on executive compensation. The literature on family firms shows that family firms run by non-professional (particularly non-founder) CEOs underperform other family firms in a wide range of countries (see Anderson and Reeb, 2003, Bennedsen et al., 2007, PTrez-Gonzlez, 2006, Sraer and Thesmar, 2007 and Villalonga and Amit, 2006).6 Consistent with this poor performance, Bloom and Van Reenen (2007) show that family firms run by sons of the founder have relatively poor management practices. For these firms the appointment of a professional manager seems like a crucial instance of value creation. However, our results suggest that families, and specially those with sons involved in management and with an absent founder, have to pay a substantial wage premium in order to attract a professional manager. If they focus solely on the cost of a professional manager, then this may explain why some of them insist on keeping management within the family and continue to underperform. The rest of the paper is organized as follows. Section 2 discusses several theories that may explain the relationship between large shareholders and compensation. Section 3 describes the data in detail. Section 4 presents the regressions with the main result and several auxiliary predictions of different theories. Section 5 concludes.

The effect of large shareholders on executive compensation and other corporate policies is a topic of much debate. Our study comes to complement this literature by presenting empirical evidence on the relationship between different types of large shareholders and CEO compensation. This is an important concern for managers around the emerging and developing world given that a large fraction of firms are controlled by large shareholders. Data on managerial compensation in environments with large shareholders is very hard to get (Fan et al., 2011), so our paper provides a unique opportunity to examine this issue. We find that family ownership translates into higher CEO compensation, but not into higher compensation for other executives below the CEO. We find a premium of about 30 log points for professional CEOs working in family firms, after controlling for several individual and firm specific characteristics, and after using alternative econometric techniques. On the other hand, firms controlled by foreigners (mostly foreign corporations), the state, and shareholder coalitions do not pay a significant premium at the CEO level. We observe differences in executives’ careers in some of these other types of firms (e.g., in tenure or internal promotion), but these cannot account for the compensation differentials that we uncover. We find that observable skills (general or firm-specific) are not able to account for the family premium. Former compensation, which can capture unobservable skills, does not predict that a particular CEO will be hired by a family firm in the future. Ultimately, it is hard to tell apart what piece of the family premium corresponds to compensation for managerial skills, private benefits, career risk, or rent extraction. All of these different mechanisms may work simultaneously, or may be more important for certain firms or certain time periods than for others. We find that the family premium comes mostly from family firms with absent founders and where children of the founder are involved. We think this last result is important for the interpretation of our findings because it stresses the main difference of family firms with other forms of concentrated ownership: the involvement of an entire family, with its own family history, in the process of controlling and managing a firm. We believe that understanding how these factors determine the level and structure of executive compensation is an interesting area for future research.