Printer Friendly

What Are Boards For? Evidence from Closely Held Firms in Colombia.

What role(s) do boards of directors play? Although the topic has received much attention from academics and regulators (for reviews, see Hermalin and Weisbach, 2003; Adams, Hermalin, and Weisbach, 2010), most of this research assumes a corporation model as described by Berle and Means (1932)--one with a widely dispersed base of shareholders in which control is exercised by management. Yet a growing volume of research shows that such a model is more the exception than the norm; instead, most public corporations around the world--and nearly all privately held companies--have a controlling shareholder or group, for the most part individuals or families (La Porta, Lopez de Silanes, and Shleifer, 1999; Claessens, Djankov, and Lang, 2000; Faccio and Lang, 2002). This is true even in countries where corporate ownership is relatively more widely held, such as the United States (Holderness, 2009; Villalonga and Amit, 2009), and increasingly so: the number of public corporations in the United States has almost halved over the last 20 years--from over 8,000 in 1996 to about 4,300 in 2016 (Doidge, Karolyi, and Stulz, 2017; Grullon, Larkin, and Michaely, 2017), and a similar phenomenon has taken place in the United Kingdom (Mayer, 2013).

The fact that most of what we know about boards is based on widely held US corporations is of concern because the governance problems these companies face are fundamentally different from the problems faced by closely held or controlled companies (for a review of controlled companies, see Villalonga et al., 2015). Thus, the board's role in these companies is also likely to be different, in accordance with the problems it is--or should be--designed to solve. Moreover, as Adams et al. (2010) note, because corporations are legally required to have a board, prior research offers little insight into why companies may or may not want to have one board; they simply have no choice.

To address these shortcomings of existing research, one needs to look at the role of boards in a sample of companies for which establishing a board is voluntary--namely, noncorporations, which by definition are privately held. Yet data on private companies are notoriously hard to find, and governance data on such companies are even more rare. In this article we take advantage of one such rare data set to investigate the question of boards' role(s). Specifically, our sample comprises 55,313 firm-year observations from 21,417 closely held Colombian firms from 2007 to 2012, of which only 56% have a board. Using detailed data from a survey about these companies' governance practices, we examine why firms have boards and how the balance of power between the board and different shareholders affects the trade-offs among control, liquidity, and growth in these firms and, ultimately, their performance.

I. Role of the Board in Closely Held or Controlled Firms

The board of directors is the representative body that, on behalf of shareholders, advises and monitors the top management team to ensure that it exercises the powers that have been delegated to it by the firm's owners in the best interests of those owners. For instance, as part of its advisory role, the board reviews and approves major strategic and financing initiatives such as the company's strategic plan, mergers and acquisitions, significant investment projects, and new equity issuances. As part of its monitoring role, the board reviews and approves the annual operating budget or any substantive deviations from it, as well as the compensation for top management, whom it is also responsible for hiring and firing.

Although both the need for strategic advice and the need for monitoring are likely to vary across firms for multiple reasons, it is the need for monitoring that is most likely to differ systematically between widely held corporations and closely held firms because of the differences between them in the balance of power among the board, controlling shareholders, minority shareholders, and managers.

In widely held corporations, well-structured and well-performing boards should mitigate the conflict of interest between owners and managers that results from the separation of ownership and control denounced by Berle and Means (1932) and Jensen and Meckling (1976): when a firm is not managed directly by its owners but by managers hired to act as agents on their behalf, managers are likely to pursue their own interests, which may differ from those of the principals. Furthermore, ownership dispersion creates a problem of collective action where individual investors lack the power, the incentives, and, often, the information to engage in direct monitoring. In such a context, it is typically more efficient to concentrate information and authority in a representative governance body such as the board (Bainbridge, 2003).

In contrast, in closely held or controlled companies, the agency problem between owners and managers is inherently alleviated by the concentration of ownership that characterizes these firms, either by ensuring better supervision of managers by owners who have both the power and the incentives to do so, or by reuniting ownership and management within the same person or team, as is often the case (La Porta et al., 1999; Villalonga and Amit, 2006, 2010).

However, these companies typically face a different agency problem: the conflicts of interest between large, controlling shareholders and small(er), minority shareholders. (1) An example of such conflicts comes from the trade-offs between control, liquidity, and growth that closely held firms often face. Controlling shareholders are typically keen on retaining their unique position in their firms, as control is economically valuable. Moreover, when the controlling shareholder is the firm's founding family, the economic motives for retaining control are often compounded by psychological motives such as the desire to perpetuate the founder's legacy, the pride in having a family member run the business, or the emotional attachment to the company that founding families often exhibit. In fact, the empirical evidence about family firms shows that these firms have significantly lower leverage than their non-family peers, which suggests that families are reluctant to share control not only with other shareholders but also with creditors (Villalonga and Amit, 2006; Gonzalez et al., 2013). This implies that family firms often restrict themselves to retained earnings as their primary, or even only, source of financing. Thus, control typically comes at the expense of reduced liquidity and/or firm growth.

Conversely, minority shareholders, having no control rights over the firm, derive all their ownership benefits from their cash-flow rights, in the form of dividends and/or capital gains, and would therefore rather not compromise their liquidity and/or growth objectives. Because controlling owners have the upper hand, their objectives tend to prevail over those of minority shareholders, as suggested by the evidence that family firms tend to be smaller and exhibit lower growth, investment, and dividend payout ratios than their peers (Villalonga and Amit, 2006, 2010; Gonzalez et al., 2014).

Large shareholders can thus act as a double-edged sword for minority shareholders in the same firm, and boards, in turn, may either mitigate or exacerbate the power of large shareholders, depending on whose interests they truly represent and how effective they are in the execution of their functions.

A few empirical studies have examined the role of boards in publicly listed family-controlled firms. Anderson and Reeb (2004) find that board composition in family firms in the S&P 500 plays a critical role in the relation between family ownership and performance. They find that family firms outperform non-family firms when independent directors balance family board representation, but otherwise underperform. They conclude that founding families monitor the firm, while independent directors monitor the family. Conversely, Klein, Shapiro, and Young (2005) find board independence to be negatively associated with family firms' performance in a sample of Canadian firms. Barontini and Caprio (2006) find that descendant directors have a positive impact on firm valuation and operating performance. Andres (2008) finds that family firms perform better only when the founding family is still active on the board. Li and Srinivasan (2011) find that companies in which the founder serves as a director with a nonfounder chief executive officer (CEO) provide higher-powered incentives than the average US board. Gonzalez et al. (2012) find that both family directors and outside directors have a positive influence on firm performance in Colombia. Gonzalez et al. (2015) find that boards dominated by families lead to longer CEO tenures, but CEO turnover shows greater sensitivity to performance, which suggests better supervision of management when family members serve on the board.

There is even less empirical evidence about the role of boards in privately held firms, despite the fact that they constitute the majority of firms even in developed economies. The only study on the subject of which we are aware is Bennedsen (2002), who uses a sample of Danish closely held firms for which the establishment of a board is voluntary to investigate why firms have boards. He finds that the likelihood of having a board increases with firm size (which he interprets as evidence of a "governance" motive--monitoring management as a way to address the conflict of interest between owners and managers) and with the number of shareholders and ownership dispersion (which he interprets as evidence of a "distributive" motive--mitigating conflicts of interest between controlling and noncontrolling shareholders). However, he does not test for either conflict of interest or the board's actual role in resolving conflicts, and he finds no evidence that boards matter for firm performance. We seek to fill this gap by examining not only the ex ante reasons why closely held firms have boards, but also the actual roles played by boards ex post, and how these roles relate to the trade-offs between control, liquidity, and growth and, ultimately, firm performance.

II. Data and Variables

A. Sample

Our sample is based on a uniquely detailed panel data set for closely held companies that combines survey responses about each firm's corporate governance practices with shareholder-level ownership information and firm-level financial data. Information is drawn largely from one Colombian government agency, the Superintendence of Companies (Superintendencia de Sociedades [SSOC]). As part of its mandate to supervise corporate restructuring and bankruptcy processes for small, medium-sized, and large privately owned firms (to protect investors' rights and firms' continuity), the SSOC maintains financial records for all these firms. Notes to financial statements are subject to statistical confidentiality and include 16 appendices per company. These list major shareholders, appointments to the board, the CEO, auditing firms, and parent-subsidiary commercial relations.

Since 2007, the SSOC has being tracking the corporate governance practices of all firms under its supervision through the Colombian Corporate Governance and Social Responsibility Survey (hereafter, Survey). The purpose of the Survey was to facilitate early detection of firms' financial distress, whose purely quantitative predictors proved insufficient until 2006.

The Survey was administered for the first time in 2007, offering respondent firms an assessment report about their corporate governance practices. Although completing the Survey was voluntary, the SSOC did not explicitly note so, which might explain the unusually high response rate for this type of survey (65%). In 2013, the SSOC made the Survey mandatory for the largest firms under its surveillance (approximately 6,000), at which point it ceased to collect Survey information for all other firms. Therefore, 2007-2012 offers an exceptional window of opportunity to study corporate governance practices in a large and random sample of closely held firms of all sizes.

Our sample is constructed as follows. Of the 29,913 companies whose financial records were collected by the SSOC in any year during 2007-2012, we eliminate 8,496 firms that did not complete the survey in any of the six years. This leaves a sample of 21,417 firms and 55,313 firm-year observations, which represent 73% of the total asset value of firms that report financial information to the SSOC.

Table I shows the distribution of our sample across different industrial sectors (Panel A) and legal forms of organization (Panel B). As shown on Panel A, the majority of the firms are in wholesale and retail trade (32%), manufacturing (21%), or real estate (17%). Investment firms (4%) are not financial firms but investment vehicles used by large shareholders to hold their equity interests in operating companies. Hence, the sample does not include firms subject to special regulation, such as retail banks, other financial institutions, or utilities.

Panel B of Table I shows that 94% of the sample firms have a legal form of organization that affords limited liability to their owners. These include: Sociedades Anonimas (corporations), the only form of business organization that by law must have a board of directors (43% of the sample); Sociedades Limitadas (equivalent to limited liability companies in the United States) (40% of the sample); Sociedades por Acciones Simplificadas (simplified joint stock companies, a legal form of organization established in 2008 that offers less demanding requirements for incorporation and administration) (11% of the sample); and Sociedades Unipersonales (single-person companies, which unlike sole proprietorships in the United States, offer limited liability to their owners) (0.5% of the sample). Although the last three types of company are not legally required to have a board, they can and, as we later report, sometimes do.

The remaining 6% of the sample comprises three types of partnerships in which some or all of the partners (i.e., general partners) have unlimited liability: Sociedades en Comandita Simple (limited partnerships, which include both general and limited partners) (3.8% of the sample), Sociedades en Comandita por Acciones (limited joint stock partnerships, a variant of limited partnerships whose capital is divided into shares instead of units) (2.1% of the sample), and Sociedades Colectivas (general partnerships) (0.03% of the sample).

B. Variables

The Survey includes questions related to governance: legal requirements, corporate structure, transparency and disclosure, anticorruption, and social responsibility, among others. We use approximately 50 questions per year to construct the indexes we need to analyze the balance of power among controlling shareholders, minority shareholders, and boards. To this end, we classify the questions into six groups: 1) board independence, 2) board roles, 3) board functioning, 4) annual general meeting effectiveness, 5) information disclosure, and 6) top management team recruitment and compensation. Table A1 in the Appendix reproduces the survey questions we use to build each index.

In addition, we have data on the equity ownership stakes of the five largest shareholders in each firm, which we use to construct our measures of controlling shareholder(s)' power. These five largest shareholders (each of which can be either an individual or a legal entity) hold a combined average stake of 96% in their firms, and a median of 100%. Therefore, by focusing on these shareholders and the distribution of equity among them we are able to get an almost complete picture of the balance of power among controlling and minority shareholders in our sample firms.

Table II provides summary descriptions of all our variables, which are explained in greater detail in the following subsections.

1. Board-Related Measures

Because not all firms in the sample have a board of directors, we first construct Board Dummy, which is equal to one if the firm has a board, and zero otherwise. We then build different indexes related to the board. Board independence can be particularly important for the effective functioning of this governance body in closely held firms; hence, we construct a measure of the power or discretion that board members can achieve, depending on how the board is shaped. Board Independence Index is the percentage of affirmative answers out of the total number of answers provided to the four questions in the Survey that relate to directors' expertise, independence, and limitations to serving on other boards in similar industries.

The category of survey items related to the board's role includes five questions designed to assess the board's involvement in relevant advisory or supervisory tasks, as well as the process through which this involvement becomes effective. These include, among others, the way in (or extent to) which the board supervises and audits the firm's financial statements, the process related to the approval of the strategic plan, and its monitoring and compliance procedures, and how the board reports to shareholders regarding the accomplishment of strategic goals.

Taking into account that the board's role entails monitoring and advising managers, we construct a general index, Board Role Index, as the percentage of affirmative answers out of the total number of answers provided to the five questions in the Survey that relate to the tasks that boards can be expected to perform along these two dimensions. We also decompose this index into two subindexes, one for each of these roles: Board Supervision Subindex, which is the percentage of affirmative answers out of the total number of answers provided to the two questions related to monitoring tasks, and Board Strategic Involvement Subindex, which is the percentage of affirmative answers out of the total number of answers provided to the three questions in the Survey that relate to the board's involvement in defining the firm's strategy and guiding its execution.

We also take into consideration a final set of questions related to the board's functioning. Specifically, these questions allow us to identify whether board meetings have a predefined detailed agenda; whether there are established procedures to assess the performance of the board as a whole and of each director individually, including their attendance; and whether assessments are carried out by independent professionals. In addition, this set of questions reveals the existence of board committees for the audit, nomination and compensation, and corporate governance functions. Board Functioning Index is the percentage of affirmative answers out of the total number of answers given to the 12 questions regarding the functioning of the board.

2. Measures of Controlling Shareholders' Power and Identity

Our primary measure of controlling shareholders' power is Herfindahl Ownership Concentration Index based on the five largest shareholders' equity stakes, following Demsetz and Lehn (1985) and many subsequent studies of ownership structure. (2) Because the main governance problem of closely held firms, and thus the subject of our interest, is the agency conflict between controlling and minority shareholders, we do not aggregate the equity stakes of the largest shareholders, even when these are clearly family related. (3) As a result, our measures can be considered a lower bound of controlling shareholders' power or ownership concentration in the firm. Conversely, to the extent that most minority shareholders in our sample are likely to be family related (Gonzalez et al., 2014), our article sheds new light on the family firms literature by capturing within-family dynamics across shareholders.

We also use two dummy variables to account for the largest shareholder's involvement in management and/or the board: Largest Shareholder CEO Duality, which equals one when the largest shareholder serves as the firm's CEO, and Largest Shareholder Director Duality, which equals one when the largest shareholder has a seat on the board. Both kinds of involvement increase the largest shareholder's power in the firm.

In addition, we include Number of Shareholders and several variables related to the identity of controlling shareholders. Specifically, we use dummy variables to measure family control of the firm: Family Firm (defined as being under family control in a financial or managerial sense), First-Generation Family Firm, and Second- or Later-Generation Family Firm.

3. Measures of Minority Shareholders' Power

We construct several variables to capture minority shareholders' power. Nonlargest Shareholder CEO Duality is a dummy variable that equals one when one of the five largest shareholders other than the very largest is the CEO, and Nonlargest Shareholder Directors Dummy equals one when one or more such shareholders are board members. These shareholders' involvement in management and/or the board effectively turns them into insiders, granting them more power and information and hence greater ability to contest the largest shareholder's power.

In addition, we use the responses to the Survey questions related to shareholder meetings dynamics to build an Annual General Meeting (AGM) Index as a proxy for the power granted to minority shareholders by corporate governance practices at the firm level. The annual general meeting is conceived as a legitimate space for addressing conflicts of interest between controlling and minority shareholders, and between shareholders and board members. Hence, governance practices at these meetings can be a proxy for the level of minority shareholder power. The Survey informs about process aspects such as the manner in which meetings take place, the call and attendance process, the existence of a functioning manual for the meeting different from the firm's charter or by-laws, a predefined detailed agenda, the process for the approval of dividend payout, proxy voting procedures for the meeting, and approval of the minutes of the meeting, among other requirements. AGM Index is the percentage of affirmative answers out of the total number of answers given to the 15 Survey questions related to the annual general meeting.

We also calculate Transparency Index as the percentage of affirmative answers out of the total number of answers given to the 14 Survey questions related to the information made available to minority shareholders, mainly paper- and Internet-based disclosures. Higher transparency reduces information asymmetries between insiders (board and management, and most likely controlling shareholders present in either group) and outside shareholders such as minority shareholders with no representation on the board. Therefore, this index can be used to proxy for the alignment of interests among these groups.

Top Management Team (TMT) Recruitment & Compensation Index is the percentage of affirmative answers out of the total number of answers given to the five questions related to top managers' recruitment and compensation policies in the Survey. A higher degree of impartiality in the selection and compensation of key executives generally implies a higher level of professionalization in the firm and can ease agency tensions between controlling and minority shareholders.

Finally, we include a measure of Foreigners' Equity Stake.

4. Dependent Variables: Leverage, Dividend Payout, and Profitability

To examine the role of boards in the balance of power between controlling and minority shareholders, we focus on the two financial policies where the conflict of interest for these two groups of shareholders is most salient: debt and dividends. We measure Leverage by the ratio of total liabilities to total assets. (4) To analyze dividend policies, we employ two variables, depending on the statistical method we use. The first variable is Dividend Dummy, which equals one when firms pay a dividend, and zero otherwise. The second variable is Dividend Ratio, computed as the amount of dividend paid out in a given year divided by total assets in the same year. We divide dividend payout by assets rather than by sales because the investment vehicles in our sample (i.e., pure holding companies) do not report sales but usually pay high dividends. Lipson, Maquieira, and Megginson (1998), Lee (2010), and Gonzalez et al. (2014) also employ dividend-to-assets as a dependent variable.

We use return on assets (ROA) as our primary financial performance measure. We are unable to use market-based measures of performance such as Tobin's Q because the firms in our sample are unlisted.

5. Control Variables

The financial variables described above are used as dependent variables in some regressions and as control variables in others. Other financial variables we use as controls include Firm Size, Firm Age (in logarithms), Growth Opportunities, Volatility of Operating Profits, and Need for External Funding, which we calculate following Durnev and Kim (2005). We also use Corporation Dummy as an independent variable in our probit models of the propensity of firms to have a board, and as an instrument in our two-stage regression models. Finally, all of our regressions include industry dummies for each of the industries shown on Panel A of Table I.

C. Descriptive Statistics

Table III presents a statistical summary of all variables. The average firm is 19 years old and has almost US$10 million in assets. Corporation Dummy shows that 43% of the observations belong to corporations (Sociedades Anonimas), the legal form of business organization that by law must have a board of directors. However, Board Dummy shows that 56% of the observations (31,221 of 55,313 firm-years) belong to firms that have a board of directors. As these statistics suggest, 75% of firms that have boards are legally required to do so because they are formally organized as corporations, whereas 25% (7,932 firms, or 14.34% of the whole sample) voluntarily choose to have a board even when their legal form of organization does not require it. In fact, because our sample firms are private and typically have a small number of shareholders, the choice of legal form of organization is voluntary. In the following section we investigate why firms choose to have boards when they are not legally required to do so.

Sample firms have an average of 10.8 and a median of five shareholders, and Herfindahl Ownership Concentration Index is 0.45. The largest shareholder's mean equity stake is 53%, indicating that private firms in Colombia typically have a majority shareholder. Foreigners have an average equity stake of 7%, which implies that most majority shareholders in the sample are local individuals or families. Indeed, only 13% of our sample firms have foreign shareholders, 8.1% in a minority position and 4.9% in a majority position.

Family firms that consider themselves as such--a lower bound on the number of firms that are truly controlled by individuals or families--represent 48% of the sample. Of these family firms, 79% (37% of the full sample) are still in the first (founder's) generation, and 21 % (10% of the full sample) are in the second or later generation. The largest shareholder serves as CEO in 37% of the sample but only serves as a director in 7% of firms with boards. This ownership structure, where ownership and control--and often management, too--are concentrated in a single individual or family, is consistent with what studies of public corporations have found to be the norm in most of the world, and particularly in emerging countries (e.g., La Porta et al., 1999; Claessens et al., 2000; Faccio and Lang, 2002).

Within the subsample of firms that have a board (Column 2 of Table III), Board Independence Index shows an average of 0.64, similar to the averages shown by Board Role Index (0.71), Board Supervision Subindex (0.70), and Board Strategic Involvement Subindex (0.71). These statistics suggest that directors in our sample exhibit a good level of experience, independence, and involvement in monitoring and advising activities. However, the low average value of Board Functioning Index (0.37) suggests that sample firms need to improve their practices to guarantee a better operation and performance of this governance body. Similar but lower values are observed if we exclude corporations from the subsample of firms that have a board (Column 3).

AGM Index has an average value of 0.58, suggesting there is room for improvement in corporate governance practices at annual meetings. These improvements would grant more power to minority shareholders in the governance of our sample firms. Nonlargest shareholders hold the CEO position in 24% of the sample firms, and serve as directors in 10% of the firms that have a board.

Transparency Index averages 0.42, which again suggests there is room for improvement in firms' disclosure practices to mitigate information asymmetries between insiders and outsiders in these firms. TMT Recruitment & Compensation Index averages 0.49, suggesting a medium level of formalization and professionalization in our sample.

The sample average return on asset (ROA) is 6% and the dividend ratio is 5%. 84% of our sample firms achieve (positive) profits; the average ROA for only these firms is 9%. Likewise, only 35% of the sample firms pay dividends, and the average dividend ratio for these firms is 13%. The average leverage for the sample is 48%.

III. Results

A. Why Do Firms Have a Board?

To understand why firms have boards, we perform three analyses. First, we conduct t-tests of the univariate differences in observable characteristics between firms with boards and firms without boards. Second, we analyze the question in a multivariate setting through a probit model of firms' propensity to have a board as a function of several of those characteristics. This model is estimated on the entire panel data set using a random-effects estimator. Third, we use a similar probit model to analyze within-firm switches from not having a board to having one, and vice versa.

1. Univariate Tests

Table III reports the mean and standard deviation for all variables broken down by firms with and without a board (Columns 2 and 4, respectively), as well as the statistics from t-tests of differences in means between the two groups of firms (Column 5). The table shows similar statistics for the subsample of noncorporations with a board (Column 3) and t-tests of differences in means between this group of firms and the subsample of firms without a board (Column 6).

As can be expected, firms that have a board are significantly larger (US$13.9 million vs. US$4.9 million) and older (20 years vs. 18 years) than those that do not have one. This is also the case for noncorporations with boards, although the differences are smaller (US$6.2 million in sales and 19 years old). Firms with boards also have more shareholders (16.5 vs. 3.5) with more diffuse ownership among them (0.41 vs. 0.51 Herfindahl). Noncorporations with boards are closer to firms without boards on both of these measures (4 shareholders and 0.48 Herfindahl) although the differences between the two groups are still statistically significant. The largest shareholder's average equity stake is 50% for firms with boards, 55% for the subset that are noncorporations and 57% for firms without boards, and this shareholder serves as the firm's CEO in 28%, 40%, and 48% of these three groups, respectively.

The composition of the shareholder base also differs between firms with and without boards in identity and not just in structure. In particular, families are most likely to be in control among noncorporations with boards (54%) and least likely among all firms (including corporations) with boards (45%), with an intermediate degree of prevalence among firms without boards (51%). First-generation families are more prevalent than second- or later-generation families in all three groups of firms, but relatively more so in firms without boards: the split between first- and later-generation firms is about 2.8:1 among family firms with boards (3.9:1 among noncorporations) but 5.2:1 among family firms without boards. These first-generation firms are responsible for the negative sign of the West of differences in Family Firm between firms with boards and firms without boards; second- and later-generation firms yield a positive sign on the same test regardless of whether corporations are included or excluded.

Firms with boards also seem to have a more balanced distribution of power between controlling and minority shareholders than those without boards, according to all our measures of minority shareholders' power. Those firms exhibit greater involvement of minority shareholders in management (nonlargest shareholder CEO in 27% vs. 20% of each group), higher AGM Index (0.61 vs. 0.55), Transparency Index (0.45 vs. 0.37), and TMT Recruitment & Compensation Index (0.57 vs. 0.40), and higher Foreigners' Equity Stake (0.07 vs. 0.06). These results are mostly true for noncorporations with boards; the differences are smaller but still significant, except for the nonlargest shareholder-CEO and foreign equity stakes.

2. Firms' Propensity to Have a Board

Table IV reports the core results of the random-effects probit regressions of firms' propensity to have a board of directors, using Board Dummy as the dependent variable. Firm without complete ownership and financial information for at least three consecutive years drop out of the sample for the estimation of these panel data regressions, which reduces the number of observations for these analyses to 37,630 firm-year observations or less, depending on the specification. As argued previously, the most relevant sample for this analysis is one in which firms have a choice between having a board or not. Accordingly, Columns 1-5 of Table IV focus on the subsample of firms that excludes corporations, and Columns 6 and 7 use the whole sample. Column 1 presents the basic model, which we expand in Column 2 to include Number of Shareholders. Column 3 shows the results including the Family Firm dummy variable, a model that Columns 4 and 5 expand in two directions: the model in Column 4 includes additional interaction terms between Family Firm and both Largest Shareholder CEO Duality and Nonlargest Shareholder CEO Duality, and the model in Column 5 breaks down family firms into generational categories using two dummies, First-Generation Family Firm and Second- or Later-Generation Family Firm (both of which equal zero for nonfamily firms). Columns 6 and 7 replicate the specifications in Columns 4 and 5 but using the whole sample and including the Corporation Dummy (which, as expected, has a positive and significant sign) as a control variable.

The results are generally consistent with those of the univariate analyses of Table III. Column 1 of Table IV shows that Herfindahl Ownership Concentration Index maintains the negative and significant impact on the propensity to have a board as suggested by the univariate results. Although a Herfindahl index of 1 (indicating that 100% of ownership is concentrated in a single shareholder) leads to a decrease of 7.33% (the coefficient's marginal effect) in the likelihood of having a board, a Herfindahl index of 0.2 (the theoretical minimum for our measure, which indicates that the firm's equity is divided equally among the five largest shareholders) leads to a decrease of only 1.5% in the probability of having a board. Hence, greater ownership concentration in the hands of the largest shareholder decreases the probability of having a board of directors. Columns 2-7 show that, however, when Number of Shareholders is included in the model, Herfindahl Ownership Concentration Index loses its statistical significance. Additional regressions (not reported but available upon request) yield similar results when we use Largest Shareholder's Equity Stake or Contestability of the Controlling Shareholder's Power as alternative proxies for the controlling shareholder's power (instead of the Herfindahl index). Taken together, these results suggest that the number of shareholders matters more than the size of their equity stakes, whether absolute or relative among shareholders. A larger number of shareholders can heighten agency or other tensions among them and thus raise the need for corporate governance mechanisms that can ease those tensions.

Consistent with the univariate results, when the largest shareholder serves as CEO, the probability of having a board decreases significantly, suggesting a self-reinforcing form of entrenchment that may be difficult to break. This is true regardless of whether corporations are included or excluded from the analyses. Conversely, family control significantly increases the probability of having a board for companies that actually face that choice--that is, noncorporations (Columns 3 and 4 of Table IV)--but the effect is nonsignificant when corporations are included in the analysis (Column 6). The interactions of both terms are positive and significant (Columns 4 and 6), although the coefficients are smaller than those of the shareholder-CEO main effects. In other words, the negative effect of the largest shareholder-CEO duality on the likelihood of having a board is attenuated in family firms, perhaps because the perception of potential conflicts--or the desire to avoid them--may be greater within families than between controlling and minority shareholders in other types of firms.

When family firms are broken down into first generation and second or later generation, the coefficients are positive and significant for both generational categories, and larger for later-generation firms. Again, this is true both when corporations are excluded (Column 5 of Table IV) or included (Column 7). Therefore, the coefficient for first-generation firms in the analysis that excludes corporations changes sign relative to the univariate analysis. These findings suggest that the more shareholders there are in a firm, the greater the collective action problem and/or the potential for conflicts among them, and thus the greater the perceived need for a board of directors as a way to mitigate those problems. This perception may be particularly pronounced in family firms because of their desire to maintain family unity and harmony.

The results for all of the governance indexes constructed out of the Survey are consistent across the seven regressions in Table IV. AGM Index has a positive and statistically significant impact on the propensity to have a board, which suggests that minority shareholders perceive the board as a governance mechanism that may ease agency tensions between them and insiders, including the controlling or largest shareholder and/or the CEO. Alternatively, the correlation between AGM Index and Board Dummy can be interpreted to be simply reflecting the fact that firms are more or less structured in their corporate governance across all governance bodies. Whether "more structured governance" translates into "better governance" is a separate question, which we investigate in the following section.

Consistent with the results in Table III, higher TMT Recruitment & Compensation Index is associated with a higher propensity to have a board. However, the direction of causality is not clear; firms with boards are also more likely to formalize their policies regarding key executives' hiring and compensation, particularly if they have a Nomination and Compensation Committee specifically established for such purposes.

In contrast, two measures of minority shareholders' power exhibit a negative and significant influence on the likelihood of having a board and depart from the univariate results in Table III: having the CEO position occupied by a nonlargest shareholder, and Transparency Index. Having a noncontrolling shareholder in the lead management role balances the power of ownership by the controlling shareholders. Higher information disclosure mitigates agency problems among stakeholders. Either factor leads to a lower perceived necessity by shareholders--controlling and/or minority--to establish formal corporate governance mechanisms such as a board of directors. Foreigners' Equity Stake loses its statistical significance from Table III in all seven regressions of Table IV.

Overall, Table IV shows that the drivers of firm's propensity to have a board are robust to excluding or including corporations in the sample for analysis, which suggests that companies have boards not just because they are legally required to have one, but because these boards serve some (other) purposes.

3. Within-Firm Switches in Having a Board or Not Having a Board

Table V shows the number and percentage of firms that switch from not having a board to having one, or vice versa, during our sample period. Most firms (94%) never switch, but 1,081 (or 2% of our sample firms) switch from not having a board to having one, whereas 2,051 firms (4%) switch from having a board to not having one.

The number of switches from having a board to not having one varies considerably by year: in 2008 only 1% of firms ceased having a board, whereas in 2011 and 2012, the figures are 8% and 7%, respectively. This increase can in part be explained by the introduction of a new legal form of organization in Colombia in 2009, Sociedad por Acetones Simplificada (simplified joint stock company), which offers companies much greater flexibility than the corporate form while preserving most of its advantages, and does not require a board. Table A2 in the Appendix shows, in the form of a transition matrix, the switches in legal form of organization associated with the switches in firms' having a board or not having one. As the table shows, 612 (30%) of the 2,051 board eliminations come from conversions of Sociedades Anonimas (corporations)--the only form that legally requires a board--into other organizational forms in the same year. Of those, 603 are conversions into Sociedad por Acetones Simplificada. Another 195 (9.5%) of the 2,051 board eliminations are linked to switches between legal forms of organization that do not require a board to begin with. Most board eliminations, however (1,244 or 61%), are not associated with any change in organizational form.

Likewise, only 16% (177 of 1,081) of the firms that establish a board after the start of our sample period appear to do so because of legal requirements--converting to a corporation from any other legal form of organization, most often from a Sociedad Limitada (limited liability corporation), which represents 169 of those 177 cases. The remaining 84% of firms switch voluntarily, which supports our other findings that companies choose to have boards for purposes other than sheer legal compliance.

To learn more about what drives firms' choice to have a board, we estimate two sets of probit models of firms' propensity to switch and report the results in Table VI. The first set (Columns 1 and 2) compares firms that switch to having a board (for whom the dependent variable equals one only in the year in which they switch, and zero for all other years) to firms that never had a board during our sample period (for which the dependent variable equals zero for all years). The second set (Columns 3 and 4) compares firms that switch to not having a board (for which the dependent variable equals one only in the year they switch, and zero for all other years) to firms that have a board during the entire sample period (for which the dependent variable equals zero for all years). The specifications are the same as in Columns 4 and 5 (or 6 and 7) of Table IV but measuring in first differences the corporation dummy (to control for the influence of changes in legal form of organization as discussed above) as well as the number of shareholders (because increases beyond a certain threshold require changes in the legal form of organization).

As Table VI shows, family firms are significantly more likely to switch in one direction or another, whereas the equity stake of foreign shareholders is negatively related to the propensity to switch in either direction. First-generation family firms are particularly likely to switch from not having a board to having one, which is consistent with the results in Table IV. Also consistent with the results in Table IV is the sign and significance of AGM Index on all four regressions: positive on the propensity to begin having a board and negative on the propensity to stop having one. The same is true for the corporation dummy, as expected. The remaining variables are not statistically significant in the analyses of firms' propensity to switch to having a board. However, many are significant predictors of the propensity to stop having a board, and the results are highly consistent with those in Table IV. Greater ownership concentration and having the largest or nonlargest shareholder serving as CEO increases the probability of dismantling an existing board, although the largest shareholder-CEO effect is significantly attenuated in family firms. Greater transparency makes firms more likely to stop having a board whereas higher TMT Recruitment & Compensation Index has the opposite effect.

B. Balance of Power among Controlling Shareholders, Minority Shareholders, and the Board

One important question that remains unanswered in the literature is: what is the effect of having a board on performance, and on relevant financial decisions, such as capital structure and dividend policy? As explained previously, the answer to this question has eluded prior research because most empirical studies use samples of publicly listed corporations, which are legally required to have a board. However, closely held, unlisted firms can voluntarily choose their legal form of organization and/or whether to have a board when they consider this body useful and suitable to govern the firm. Hence, taking advantage of the unique characteristics of our sample, we are able to assess the impact that the existence of a board has on leverage, payout policy, and profitability after controlling for the impact of controlling and minority shareholders on the same variables. By focusing on these three dependent variables, about which the objectives of controlling and minority shareholders may differ, we are also able to evaluate the role played by the board in balancing the relative power between these two groups of shareholders.

As before, we perform three groups of analyses to answer this question. First, we estimate single-stage panel data models on the entire sample (random effects or Tobit, according to the distribution of the dependent variable). Second, we estimate two-stage treatment-effects models to control for the endogeneity of the decision to establish a board. Third, we use similar treatment-effects models to analyze the impact on leverage, payout policy, and profitability of within-firm switches from not having a board to having one, or vice-versa.

1. Random-Effects Models of the Board's Impact on Financial Policies and Performance

Columns 1 and 3 of Table VII show the results of random-effects panel regressions that use Leverage and ROA as dependent variables, respectively. In Column 2 we estimate a similar model using Dividend Ratio as the dependent variable. Because this variable is censored at zero, however, we estimate it as a Tobit model.

The results in Columns 1 and 2 of Table VII suggest there is indeed tension between controlling and minority shareholders with respect to financial policies on leverage and dividends. More to the point of this article, they also suggest that when there is a divergence of interests between the two groups of shareholders, boards support the largest shareholder's decisions.

Regarding debt policy (Column 1 of Table VII), the positive and significant impact of Herfindahl Ownership Concentration Index on total leverage suggests a preference for higher debt levels on the part of controlling shareholders, most likely to finance growth without losing control. This preference is further reinforced--or more easily implemented--when the largest shareholder acts as the firm's CEO. In contrast, AGM Index shows the opposite effect, suggesting that minority shareholders prefer to avoid the risk and pressure that debt exerts on the firm's free cash flows, and the possible reduction in their dividends that may come with it. Foreigners' Equity Stake is also negatively associated with leverage, which is consistent with the minority position this group holds, on average, in our sample. Board Dummy has a positive and significant impact on total leverage, suggesting that the board endorses the decisions favored by controlling shareholders: having a board makes firm leverage one percentage point higher.

Two measures of minority shareholder power other than AGM Index also have a positive and significant effect on leverage, suggesting ways the different preferences of controlling and minority shareholders can be aligned. One is having greater Transparency Index, which leads to higher leverage, perhaps because greater disclosure implies lower information asymmetries between insiders and potential financial creditors, making it easier for private firms to obtain external funding. TMTRecruitment & Compensation Index also exhibits a positive and significant relation with total leverage, which suggests that greater clarity in the way key executives are hired and compensated helps align these managers' objectives with those of controlling shareholders.

Conversely, family shareholder-CEOs have a preference for lower leverage regardless of whether they are minority or controlling shareholders: the interaction coefficients are both negative and statistically significant and offset the main effects of the shareholder-CEO measures. Although this preference contrasts with that of (non-family) controlling shareholders and the board, the results are consistent with earlier findings in the family firms literature that families as controlling shareholders are significantly more averse to debt than other owners (Anderson and Reeb, 2003; Villalonga and Amit, 2006). With a different sample of private firms in Colombia, Gonzalez et al. (2013) find that family managers have a negative influence on leverage.

Regarding payout policy (Column 2 of Table VII), ownership concentration has a negative and significant effect that is consistent with controlling shareholders' preference to retain earnings to finance future growth (or debt, as shown previously) to avoid having to raise outside equity and thereby dilute the stake that entitles them to control the firm. However, this preference is attenuated (rather than exacerbated, as is the case with leverage) when the largest shareholder serves as CEO, and even more so when the firm is a family firm.

Minority shareholders exhibit the opposite preference to controlling shareholders with respect to dividends as they do with respect to leverage: three of our proxies for minority shareholder power indicate their preference for higher dividends. An increase of 0.1 in AGM Index leads to an average increase of approximately 5.7% on the dividend ratio. Using AGM Index as a proxy for minority shareholder power is of particular relevance for payout policy decisions in the context of this study because in Colombia, as in Continental Europe (but not in the United States), the power to approve the annual dividend resides with shareholders, not with the board (Cools, 2005). Thus, a more effective functioning of annual general meetings as captured by this index helps minority shareholders implement their preferences with respect to dividends. Having the CEO position also helps minority shareholders realize their preferences in this regard in a statistically significant way, as does having greater transparency.

As is the case with leverage, the board seems to align with controlling shareholders regarding dividend payments: Column 2 of Table VII shows that the dividend ratio is 6% lower in firms that have a board of directors. These results point to tension between minority shareholders and boards, whereby minority shareholders exhibit a preference for greater liquidity in the form of dividends, whereas directors favor retaining earnings to fund investments and growth opportunities.

As is the case with leverage, the coefficient of TMT Recruitment & Compensation Index in the dividend ratio regression has the same sign as the controlling shareholders' and the board's, and opposite to those of the measures of minority shareholder power. The negative and significant impact of TMT Recruitment & Compensation Index on the dividend ratio reinforces the evidence that more explicit hiring and compensation policies align the objectives of the top management team with those of controlling shareholders and of the boards that act on their behalf.

The results in Column 3 of Table VII indicate that having a board matters for firm performance, in a positive and statistically significant way. Sample firms with boards have ROAs that are, on average, 1% higher than firms without boards. This effect is significantly reinforced when the largest shareholder is represented on the board, which adds an extra 1% to ROA. In contrast, the presence of the nonlargest shareholder on the board has a negative effect on ROA that outweighs the positive effect on the board by itself (-1%). Also contrasting with the positive effect of the largest shareholder-director duality is the finding that when the largest shareholder serves as CEO, ROA is reduced by 1%. This result is consistent with the arguments and findings of managerial entrenchment in the ownership-performance literature (Morck, Shleifer, and Vishny, 1988; Stulz, 1988). On that note, it is worth highlighting that ownership concentration has no direct effect on performance, a result--or lack thereof--that is consistent with the arguments in Demsetz (1983) and the results in Demsetz and Lehn (1985) and Demsetz and Villalonga (2001). What is different and perhaps more remarkable about our finding is that in our sample, the relation proves to be insignificant even without controlling for the endogeneity of ownership structure.

Our four Survey-based measures of minority shareholders' power have positive coefficients that are significant at the 1% level. First, an increase of 0.1 in AGM Index increases accounting performance by 0.8%. As mentioned previously, the annual general meeting is conceived as a legitimate space for addressing conflicts of interest between controlling and minority shareholders, and between shareholders and board members. Hence, good practices at annual general meetings foster better governance and better performance. Second, an increase of 0.1 in Transparency Index increases ROA by 1.5%, suggesting that disclosure practices that reduce information asymmetries between insiders and outsiders are beneficial for all capital providers. This finding is consistent with prior studies, which find positive performance effects of disclosure on firm accounting and market performance in emerging markets (e.g., Garay et al., 2013). Third, an increase of 0.1 in TMT Recruitment & Compensation Index also increases ROA significantly, albeit minimally (by 1%). Finally, Foreigners' Equity Stake has a coefficient of 4%, suggesting that international shareholders introduce discipline or management practices that translate into higher performance.

Taken together, the positive and significant coefficients of these four measures of minority shareholder power, as well as that of Board Dummy, suggest that more structured governance translates into better governance, as evidenced by firm performance.

2. Treatment Effects of the Board on Financial Policies and Performance

We estimate treatment-effects models to control for the endogeneity of the decision to have a board when assessing the effect of having a board on performance and on the financial policy outcome variables. Columns 4-6 of Table VII present the results of the second stage in two-step treatment-effects regressions where Board Dummy serves as the treatment indicator, and the first-stage regression follows the specification in Column 6 of Table IV (the most complete among those estimated on the full sample).

We use Corporation Dummy as an instrumental variable to estimate the model via Heckman's procedure. As shown in Table IV, this legal form of organization is an important determinant of a firm's decision to have a board, as no other form is legally required to have one. (5) Whether the choice of a corporate form affects financial policies and performance--and hence, whether our proposed instrument meets the exclusion restriction--depends largely on the liability protection, access to capital, and tax treatment offered by alternative organizational forms. With unlimited liability, firm owners are more reluctant to assume debt and to pay dividends, as they have to rely to a greater extent on retained earnings to finance their firms' growth. Ultimately, growth and profitability may be compromised as well.

Debt and profitability may also be affected by whether a company is considered a "pass-through" entity for tax purposes (i.e., the company as such is not taxed and instead passes its income to its shareholders, who are taxed at their personal tax income rates); if personal and company tax rates differ substantially, the incentives to take on debt or book greater profits may differ systematically between pass-through entities and other types of company.

Leaving partnerships aside, there are no meaningful differences in our sample across the remaining legal forms of organization in the access to capital that each may provide. Although, as in other countries, the corporate form (Sociedad Anonima) is required for firms to access public stock markets, in Colombia this market is of little relevance: the number of listed firms is very small, and like in the United States, it has been declining over the past two decades, including our sample period: from 96 in 2007 to 76 in 2012. These figures are also much lower than those for Brazil, Chile, Peru, Mexico, and Argentina over the same period. Between 1990 and 2014, there were only 15 initial public offerings (IPOs). Therefore, it seems extremely unlikely for companies in our sample to choose to organize themselves as corporations to go public. Moreover, limited liability companies in Colombia are allowed to have up to 25 (private) shareholders, a constraint that is binding for very few firms in our sample firms: only 5% of all corporations have more than 25 shareholders. Therefore, access to capital does not seem to drive the choice of legal form of organization by the nonpartnership firms in our sample.

To ensure the validity of our instrument, we therefore eliminate from the sample all partnerships (general or limited), which as shown in Table I, represent less than 6% of the full sample. All remaining firms have limited liability and are not pass-through entities regardless of their legal form of organization. This elimination does not materially affect the results of our treatment-effects regressions but removes any statistically significant differences between corporations and other companies in their leverage, dividend ratio, or profitability, thus validating our instrument empirically as well as theoretically. The results of these tests are not reported here but are available from the authors upon request.

Columns 4-6 of Table VII show the second-stage results on this restricted sample using Leverage, Dividend Dummy, and ROA as dependent variables, respectively. We use a dividend dummy instead of the dividend ratio used previously because the latter variable is censored but the Tobit model cannot be estimated as the second stage of our treatment effects regression.

Overall, the results of the single-stage regressions in Columns 1-3 of Table VII are maintained or strengthened in the two-stage models. All the variables that were statistically significant keep their sign, and the coefficients become larger. As a result, the statistical significance is the same or greater, with the exception of the two shareholder-director duality measures in the ROA regression, which lose their statistical significance despite their coefficients becoming larger. Furthermore, several variables that were not significant in the single-stage regressions become statistically significant after the endogeneity of the decision to have a board is controlled for.

Most notably, a nonlargest shareholder serving as CEO or as a director on the board now has positive and significant effects on leverage. The implication is that nonlargest shareholders collude with the largest shareholders after they become part of the board or the management team, thereby exacerbating the divergence of interests between the board and minority shareholders that are not represented in either group. Taken together, the findings that both the largest and nonlargest shareholders acting as CEOs increase leverage as well as the likelihood of distributing dividends, though not intuitive, are consistent with the predictions of Zwiebel's (1996) model, in which managers voluntarily issue debt and pay dividends to avert challenges to their control.

The treatment-effects models also highlight the role of families as shareholders. In the single-stage models in Columns 1-3 of Table VII, family firms exhibit a distinct impact on financial policies only when the largest or nonlargest shareholder is the CEO. In contrast, Column 5 shows that family firms are more likely to pay dividends, which partially offsets the negative coefficient of ownership concentration. This result, though consistent with earlier findings for Colombia (Gonzalez et al., 2014), runs contrary to what has been found in other parts of the world, such as the United States (Villalonga and Amit, 2006). One possible explanation for this discrepancy in results arises from the fact that studies from other parts of the world are based on publicly listed firms, in which minority shareholders can obtain liquidity by selling their shares on the public market. Thus, they are less likely to press for dividends than when they have no other means to obtain liquidity, as is typically the case in private firms. Moreover, because studies of family firms do not distinguish between the objectives of controlling and minority shareholders (family or non-family), it is possible that minority shareholders may also press for dividends in those firms, but ultimately it is the preference of large shareholders for reinvestment that prevails.

Family Firm also has a positive and significant effect on performance in the treatment-effects regression in Column 6 of Table VII. This finding is consistent with what the now voluminous literature on family firms' performance has shown to be a largely universal result (for a review, see Amit and Villalonga, 2014).

Altogether, our treatment-effects analyses show that our findings are highly robust to controlling for the endogeneity of the decision to have a board when assessing the board's impact on financial policies and performance.

3. Treatment Effects of Switching to Having a Board or Not Having a Board on Financial Policies and Performance

To take further advantage of the panel nature of our data set, we analyze how financial policies and performance change following firm switches from not having a board to having one, or vice versa. Specifically, we estimate treatment-effects models similar to those in the last three columns of Table VII but measuring the dependent variables in changes over the year following the switch instead of levels, and replacing the board dummy with two board switch dummies (one for each direction), which equal one in the firm and year of the switch, and zero otherwise. Our expectation is therefore that the dummy that indicates a firm's switch from "without a board" to "with a board" will have regression coefficients of the same sign as those in Table VII, whereas the dummy that indicates a firm's switch from "with a board" to "without a board" will yield regression coefficients with opposite signs. All other independent variables are expected to maintain the signs in Table VII.

Table VIII shows the results of these analyses. Consistent with the results of our analyses of levels, firms experience significant increases in leverage and decreases in dividend payments after switching to having a board and, conversely, significant decreases in leverage after switching to not having a board. Board eliminations have no significant effect on dividends, however, perhaps because dividends are stickier over time than debt (in general but particularly so in Colombia, where there are certain minimum thresholds established by law).

Firms also experience significant decreases in ROA following board switches in either direction. The decline in performance following board eliminations is consistent with the finding of a positive effect of boards on performance levels. The decline in performance following board initiations, though harder to reconcile with the results of the analysis of levels, suggests an interesting dynamic that includes reverse causation: an anticipated decline in performance in a given year may lead companies that do not have a board to establish one as a form of strategic intervention. Once the board is established, it helps firms overcome their difficulties, which leads to the more general result across the entire panel data set of a positive impact of boards on financial performance.

Besides the effects of board switches just reported, all independent variables in Table VIII maintain the signs and, for the most part, the statistical significance they had in Table VII. This shows that our results are robust not only to endogeneity concerns but also to whether the effects are measured in levels or in changes.

Overall, boards seem to support large shareholders' decisions regarding both debt and dividends. The implication is that when controlling and minority shareholders have divergent interests, boards act on behalf of the controlling shareholders. This is perhaps not surprising, given that it is these shareholders who effectively delegate to the board some of the governance and management functions that are associated with control rights, who are likely to have a greater say in the appointment of directors to the board, and therefore, whom directors are more likely to truly represent.

C. Board's Roles and Effective Functioning

To investigate the strategic and monitoring roles played by the board in companies that choose to have one, as well as the board's effective functioning, we use Heckman's two-step approach again to estimate sample selection models. Table IX presents the results of the second-stage regressions for these specifications. Like the treatment-effects models of Table VII, these selection models use Board Dummy as the selection indicator, the first-stage regression follows the specification in Column 9 of Table IV, and we apply the exclusion restriction to Corporation Dummy. Columns 1 and 2 of Table IX use Total Leverage, Columns 3 and 4 use Dividend Dummy, and Columns 5 and 6 use ROA as the dependent variables for the second-stage regressions. Odd columns use Board Role Index as the explanatory variable, and even columns use two subindexes: Board Supervision Subindex and Board Strategic Involvement Subindex.

The results of these estimations show that the board's dual roles in monitoring and advising managers, and its effective functioning, have significant effects on leverage, the probability of dividend payments, and firm performance.

Board Role Index has a positive and significant effect on total leverage (Column 1 of Table IX). The breakdown of Board Role Index into its subcomponents, Board Supervision Subindex and Board Strategic Involvement Subindex (Column 2), shows that this effect is entirely attributable to the board's involvement in formulating the firm's strategy and supervising its implementation. Board Strategic Involvement Subindex exhibits a positive and significant coefficient on leverage, whereas Board Supervision Subindex has no significant effect. This result is consistent with the idea that boards favor a higher level of debt to fund strategic investments and growth opportunities.

The same reasoning applies to the negative and significant effects that Board Role Index and Board Strategic Involvement Subindex have on the probability of dividend payments (Columns 3 and 4 of Table IX). In the case of dividends, these effects are further reinforced by the negative and significant effects of Board Supervision Subindex, Board Independence Index, and Board Functioning Index. Boards seem to have a preference to use retained earnings to finance growth, and the better they work, the more effective they are in implementing this preference. These results are consistent with the idea that boards support controlling shareholders' decisions regarding leverage and dividend payout policies, and cast doubt on the real independence of the boards in our sample. In private, closely held firms, controlling shareholders have enough power to determine board composition and to appoint "gray" directors as independent when they are not completely so.

Both Board Role Index and Board Functioning Index are also negatively and significantly related to firm performance (Column 5 of Table IX). The breakdown of Board Role Index into its subcomponents (Column 6) shows that the negative sign on this variable is again attributable to both of its subindexes: Board Supervision Subindex and Board Strategic Involvement Subindex.

These findings suggest a likely reverse causation from firm performance to boards' role and functioning that is consistent with the negative effect of board initiations reported earlier: boards may have to play a more active and strategic role when companies underperform.

The rest of the variables exhibit coefficients that are largely consistent with the results discussed in Table VII, although the significance levels vary (typically in the direction of higher significance in Table IX). We perform additional robustness checks to check our results, which are not reported here but are available from the authors upon request. First, we run our regressions without considering the investment vehicles used by the largest shareholders to hold their equity interests in operating companies. Second, we rerun our regressions without including firms whose boards have less than six members. In all of these robustness checks, our results remain practically unchanged.

IV. Conclusion

In this article, we use a large sample of closely held firms to investigate the role of the board and the drivers of the decision to have one. We examine the role played by the board in the balance of power between controlling and minority shareholders by examining the impact that each of these three groups has on performance, leverage, and payout policy.

We find that the number and identity of shareholders are more important determinants of firms' propensity to have a board of directors than the size of their equity stakes. As can be expected, the larger the number of shareholders, the greater the probability of their having a board of directors to represent them. Being a family firm also increases the probability of having a board, especially when the firm is in its second and later generation.

These findings suggest that the larger the number of shareholders, the greater the collective action problem and/or potential for conflicts, and thus the greater the perceived need for a board of directors as a way to mitigate these problems. This perception seems particularly pronounced among family shareholders, perhaps driven by their concern for maintaining unity and harmony in the family. Indeed, after controlling for the endogeneity of the decision to establish a board, we find that all shareholders benefit from having a board in the form of higher profitability.

Yet we also find that when the preferences of controlling and minority shareholders diverge, boards align with controlling shareholders, as can be expected because it is this group who effectively delegates to the board some of the valuable decision rights that come with their controlling position. Specifically, boards and controlling shareholders favor greater leverage and lower dividends, which help these shareholders preserve their controlling equity stake, whereas minority shareholders favor lower leverage and higher dividends, which provide them with a safer return on their investment.

Altogether, our results suggest that boards play a different role in closely held and controlled firms from what the voluminous literature about boards has shown for widely held corporations. In closely held firms, where the dominant agency problem is between controlling and minority shareholders, and not between owners and managers, boards appear to side with controlling shareholders, who effectively delegate their functions in the board and management. As a result, boards sharpen both edges of the double-edged sword that large shareholders represent for minority shareholders in the same firm: although they help controlling shareholders monitor managers for the collective benefit of all shareholders, they also facilitate the appropriation of private benefits of control by large shareholders. This finding has important managerial and policy implications regarding closely held and controlled firms, which constitute a growing majority of firms around the world.

Appendix: Corporate Governance in Colombia

Colombia presents similar characteristics in terms of corporate governance to other emerging markets and even to some developed economies, including the United States. La Porta et al. (1998) find high levels of ownership concentration for Colombian firms (combined ownership of 63% by the three largest shareholders), similar to other developed economies such as Belgium, Italy, Portugal, and Greece, and also to emerging markets like Brazil, Egypt, Indonesia, and Mexico. According to La Porta et al. (1997), Colombia achieves only one of five possible points in its anti-director rights index alongside with Greece, Germany, and Uruguay. The score for the rule of law in the same study for Colombia is only 2.08, similar to Peru, and the Philippines. In the same way, Colombia gets 0 of 4 possible points in its creditor rights index, similar to France, Peru, United States, UK, and Canada. External capitalization to the country gross nation product (GNP) in La Porta et al. (1997) is only 0.14 for Colombia, comparable to Argentina, Belgium, Italy, Portugal, Spain, Turkey, and Germany.

In a more recent study, Chong and Lopez de Silanes (2007) argue that capital markets in Latin America are not a real source of funding, among other reasons due to the poor protection they afford investors. Their statement finds support in 2012 World Bank data, which show that local listed firms are only 1,208 in all Latin America: 353 in Brazil, 225 in Chile, 213 in Peru, 131 in Mexico, 101 in Argentina, and 76 in Colombia. These numbers imply a low level of listed firms per capita (million): 2.08 across all of Latin America, or 2.46 for Argentina, 1.78 for Brazil, 1.59 for Colombia, and 1.08 for Mexico, for instance.

With fewer than 100 firms listed, the vast majority of Colombia's 1.6 million firms are closely held, unlisted firms. The importance of closely held firms relative to public corporations is not unique to Colombia and other emerging economies, however, but also characteristic of many developed economies including--and increasingly so--the United States'. The World Bank's 2017 Doing Business survey highlights the existence of 5.9 million firms in US, of which only about 4,300 were listed--down from over 8,000 in 1996 (Doidge et al., 2017; Grullon et al., 2017), Doidge et al. (2017) rank Colombia fourth, and the United States seventh, out of 54 countries in the decrease in the number of domestic, publicly listed firms between 1996 and 2012. Doidge et al. (2017) also show that the decrease results from a combination of factors consistent with the decrease of the net benefits of being publicly listed in the United States: a decline in IPOs as first documented by Gao, Ritter, and Zhu (2013), and an increase in delistings that is largely attributable to acquisition of public companies (Grullon et al., 2017). These trends and figures highlight the relevance of analyzing corporate governance mechanisms and practices in closely held firms.

With a gross domestic product (GDP) of $688.8 billion, Colombia is the fourth largest economy in Latin America (after Brazil, Mexico, and Argentina), and the thirty-second in the world. By way of comparison, the United States ranks third, with a GDP of $ 18.57 trillion. The United States also has the third largest population in the world, with approximately 327 million inhabitants; Colombia ranks twenty-ninth with approximately 48 million. As a result, Colombia's GDP per capita (US$14.1) is approximately a quarter of the United States'. According to the World Bank's 2017 Doing Business survey, Colombia has the second best business environment in Latin America, after Mexico.

In recent years, Colombia has implemented significant changes in corporate governance. Corporate regulations have been partially reformed, and corporate governance country codes were created in 2007 for listed firms and in 2009 for closely held firms. The corporate governance code for listed firms was updated in 2014 to include more demanding recommendations. Colombia scored three of five possible points in the revised anti-director rights index of Djankov et al. (2008), and by our own estimations, the score would now be equal to four if the current regulatory framework were taken into account. Djankov et al. (2006) also note that Colombia is among the lower-middle-income countries that do quite well on debt enforcement. The main weakness for Colombia remains in public enforcement (Djankov et al., 2008).

Although government securities continue to dominate the local capital market, stocks have increased their share of total volume traded. According to the Colombian Stock Exchange, in 2002 government securities represented 84% of the total volume traded, corporate bonds 15%, and stocks 1%. By 2014 these percentages had changed to 86%, 6%, and 8%, respectively. External capitalization relative to the country's GDP was 71% in the World Bank's latest statistics (2012). However, only about 30 of the 76 firms listed in Colombia have highly traded shares.

In Colombia shareholders enjoy substantial power, and even more in closely held, unlisted firms. By law, as in Continental Europe, shareholder meetings approve dividends, and shareholders can appoint and remove directors at any time. Moreover, only corporations (sociedades anonimas), which represent less than the 40% of all firms in Colombia, are legally required to establish a board of directors. Hence, the majority of closely held, unlisted firms voluntarily constitute boards when they consider this corporate governance body valuable. Furthermore, only corporations must have an independent external financial audit. Legal requirements regarding the level of disclosure and other corporate governance practices, and the intensity of the supervision exerted by governmental surveillance bodies, are lower for closely held, unlisted firms than for publicly listed corporations.

When it is mandatory for firms to have a board, this corporate governance body must be comprised of a minimum of three members, each of which has to have an "alternate" nominated. Firms that voluntarily constitute a board are not required to have a minimum number of directors. However, in our sample there are no single-member boards, and only 1.94% of all boards have two members (plus their respective alternates). Hence, the vast majority of boards in Colombia have three or more members (plus, again, their alternates).

Finally, pyramidal control and cross-shareholdings are common among large Colombian firms, most of which have business group affiliation status (an official designation), but not among other firms. The 25 largest Colombian family business groups altogether have less than 500 firms under their control (Gonzalez et al., 2014, 2015). The average firm in Colombia is subject to the direct control of its founding family shareholders, the majority of which are individuals. As a result, in our sample, although there are some investment vehicles used by largest shareholders to hold their equity interests in operating companies (4% of all sample firms), they are no representative pyramidal or cross-shareholding structures.

References

Adams, R., B. Hermalin, and M. Weisbach, 2010, "The Role of Boards of Directors in Corporate Governance: A Conceptual Framework and Survey," Journal of Economic Literature 48, 58-107.

Amit, R. and B. Villalonga, 2014, "Financial Performance of Family Firms," in L. Melin, M. Nordqvist, and P. Sharma, Eds. The SAGE Handbook of Family Business, London, SAGE Publications, 157-178.

Anderson, R. and D. Reeb, 2003, "Founding-Family Ownership and Firm Performance: Evidence from the S&P 500," Journal of Finance 58, 1301-1328.

Anderson, R. and D. Reeb, 2004, "Board Composition: Balancing Family Influence in S&P 500 Firms," Administrative Science Quarterly 49, 209-237.

Andres, C, 2008, "Large Shareholders and Firm Performance: An Empirical Examination of Founding-Family Ownership," Journal of Corporate Finance 14, 431-445.

Bainbridge, S., 2003, "Director Primacy: The Means and Ends of Corporate Governance," Northwestern University Law Review 97, 547-606.

Barontini, R. and L. Caprio, 2006, "The Effect of Family Control on Firm Value and Performance: Evidence from Continental Europe," European Financial Management 12, 689-723.

Bennedsen, M., 2002, "Why Do Firms Have Boards?" SSRN Working Paper.

Bennedsen, M. and D. Wolfenzon, 2000, "The Balance of Power in Closely Held Corporations," Journal of Financial Economics 58, 113-139.

Berle, A. and G. Means, 1932, The Modern Corporation and Private Property, New York, Transaction Publishers.

Chong, A. and F. Lopez de Silanes, 2007, "Corporate Governance in Latin America," in A. Chong and F. Lopez de Silanes, Eds. Investor Protection in Latin America, Redwood City, CA, Stanford University Press, 1-84.

Claessens, S., S. Djankov, and L. Lang, 2000, "The Separation of Ownership and Control in East Asian Corporations," Journal of Financial Economics 58, 81-112.

Cools, S., 2005, "The Real Difference in Corporate Law between the United States and Continental Europe: Distribution of Powers," Delaware Journal of Corporate Law 30, 697-766.

Demsetz, H., 1983, "The Structure of Ownership and the Theory of the Firm," Journal of Law and Economics 26, 375-390.

Demsetz, H. and K. Lehn, 1985, "The Structure of Corporate Ownership: Causes and Consequences," Journal of Political Economy 93, 1155-1177.

Demsetz, H. and B. Villalonga, 2001, "Ownership Structure and Corporate Performance," Journal of Corporate Finance 7, 209-233.

Djankov, S., O. Hart. C. McLiesh, and A. Shleifer, 2006, "Debt Enforcement around the World," Journal of Political Economy 116, 1105-1149.

Djankov, S., R. La Porta, F. Lopez de Silanes, and A. Shleifer, 2008, "The Law and Economics of Self-Dealing," Journal of Financial Economics 88, 430-465.

Doidge, C.G., A. Karolyi, and R. Stulz, 2017, "The US Listing Gap," Journal of Financial Economics 123, 464-487.

Durnev, A. and H. Kim, 2005, "To Steal or Not to Steal: Firm Attributes, Legal Environment, and Valuation," Journal of Finance 60, 1461-1493.

Faccio, M. and L. Lang, 2002, "The Ultimate Ownership of Western European Corporations," Journal of Financial Economics 65, 365-395.

Gao, X., J.R. Ritter, and Z. Zhu, 2013, "Where Have All the IPOs Gone?" Journal of Financial and Quantitative Analysis 48, 1663-1692.

Garay, U., M. Gonzalez, A. Guzman, and M.-A. Trujillo, 2013, "Internet-Based Corporate Disclosure and Market Value: Evidence from Latin America," Emerging Markets Review 17, 150-168.

Gonzalez, M., A. Guzman, C. Pombo, and M.-A. Trujillo, 2012, "Family Firms and Financial Performance: The Cost of Growing," Emerging Markets Review 13, 626-649.

Gonzalez, M., A. Guzman, C. Pombo, and M.-A. Trujillo, 2013, "Family Firms and Debt: Risk Aversion versus Risk of Losing Control," Journal of Business Research 66, 2308-2320.

Gonzalez, M., A. Guzman, C. Pombo, and M.-A. Trujillo, 2014, "Family Involvement and Dividend Policy in Closely Held Firms," Family Business Review 27, 365-385.

Gonzalez, M., A. Guzman, C. Pombo, and M.-A. Trujillo, 2015, "The Role of Family Involvement on CEO Turnover: Evidence from Colombian Family Firms," Corporate Governance: An International Review 23, 266-284.

Grullon, G., Y. Larkin, and R. Michaely, 2017, "The Disappearance of Public Firms and the Changing Nature of US Industries," Rice University Working Paper.

Hermalin, B. and M. Weisbach, 2003, "Boards of Directors as an Endogenously Determined Institution: A Survey of the Economic Literature," Federal Reserve Bank of New York Economic Policy Review 9, 7-26.

Holderness, C, 2009, "The Myth of Diffuse Ownership in the United States," Review of Financial Studies 22, 1377-1408.

Jara-Bertin, M., F. Lopez-Iturriaga. and O. Lopez de Foronda, 2008, "The Contest to the Control in European Family Firms: How Other Shareholders Affect Firm Value," Corporate Governance: An International Review 16, 146-159.

Jensen, M. and W. Meckling, 1976, "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure," Journal of Financial Economics 3, 305-360.

Klein, P, D. Shapiro, and J. Young, 2005, "Corporate Governance, Family Ownership and Firm Value: the Canadian Evidence," Corporate Governance: An International Review 13, 769-784.

La Porta, R., F. Lopez De Silanes, and A. Shleifer, 1999, "Corporate Ownership around the World," Journal of Finance 54, 471-517.

La Porta, R., F. Lopez de Silanes, A. Shleifer, and R. Vishny, 1997, "Legal Determinants of External Finance," Journal of Finance 52, 1131-1150.

La Porta, R., F. Lopez de Silanes, A. Shleifer, and R. Vishny, 1998, "Law and Finance," Journal of Political Economy 106, 1113-1155.

Lee, K, 2010, "Retail Minority Shareholders and Corporate Reputation as Determinant of Dividend Policy in Australia," Pacific-Basin Finance Journal 18, 351-368.

Li, F. and S. Srinivasan, 2011, "Corporate Governance When Founders Are Directors," Journal of Financial Economics 102, 454-469.

Lipson, M., C. Maquieira, and W. Megginson, 1998, "Dividend Initiations and Earnings Surprises," Financial Management 27, 27-45.

Maury, B. and A. Pajuste, 2005, "Multiple Large Shareholders and Firm Value," Journal of Banking & Finance 29, 1813-1834.

Mayer, C, 2013, Firm Commitment. Why the Corporation Is Failing Us and How to Restore Trust in It, Oxford, UK, Oxford University Press.

Morck, R., A. Shleifer, and R. Vishny, 1988, "Management Ownership and Market Valuation: An Empirical Analysis," Journal of Financial Economics 20, 293-315.

Stulz, R.M., 1988, "Managerial Control of Voting Rights: Financing Policies and the Market for Corporate Control," Journal of Financial Economics 20, 25-54.

Villalonga, B., and R. Amit, 2006, "How Do Family Ownership, Control, and Management Affect Firm Value?" Journal of Financial Economics 80, 385-417.

Villalonga, B. and R. Amit, 2009, "How are US Family Firms Controlled?" Review of Financial Studies 22, 3047-3091.

Villalonga, B. and R. Amit, 2010, "Family Control of Firms and Industries," Financial Management 39, 863-904.

Villalonga, B., R. Amit, M.-A. Trujillo, and A. Guzman, 2015, "Governance of Family Firms," Annual Reviews of Financial Economics 7, 635-654.

Zwiebel, J., 1996, "Dynamic Capital Structure under Managerial Entrenchment," American Economic Review 86, 1197-1215.

Belen Villalonga, Maria-Andrea Trujillo, Alexander Guzman, and Neila Caceres (*)

We thank Saverio Bozzolan, Eduardo Gentil, Han Kim, Hannes Wagner, David Yermack, an anonymous referee, and participants in seminars at Bocconi, LUISS Guido Carli, the University of Michigan, New York University, Tulane University, Washington University at St. Louis, CESA School of Business, and at the annual meetings of the Northern Finance Association 2015, Financial Management Association 2016. and Midwest Finance Association 2016 for their comments and suggestions. All errors are our own.

(*) Belen Villalonga is a Professor of Management and a Professor of Finance (by courtesy) at New York University Stern School of Business in New York. Maria-Andrea Trujillo is a Professor of Finance at the CESA School of Business in Bogota, Colombia. Alexander Guzman is a Professor of Finance at the CESA School of Business in Bogota, Colombia. Neila Caceres is a Researcher at the Superintendencia de Sociedades in Bogota, Colombia.

(1) Because, in our sample, noncontrolling shareholders are typically minority holders, in this article we use the term "minority shareholders" to refer, indistinctively, to noncontrolling shareholders and/or minority shareholders.

(2) As a robustness check, we re-estimate all of our regressions using two alternative measures: the largest shareholder's equity stake and Contestability of the Controlling Shareholder's Power. The latter measure is defined as the ratio of the combined equity stakes of the second, third, fourth, and fifth largest shareholders to the largest shareholder's equity stake. Similar measures are used or proposed in studies by Bennedsen and Wolfenzon (2000), Maury and Pajuste (2005), and Jara-Bertin, Lopez-Iturriaga, and Lopez de Foronda (2008). Because all of these measures are highly correlated, we can only use one at a time in each regression. However, the results are similar when using any of the measures and are available from the authors upon request.

(3) Ideally, we would like to examine the balance of power between controlling and minority shareholders separately for minority shareholders who are family related to the controlling shareholder and for those who are not, but given the size and privately held status of our sample firms, we are unable to verify family relationships for most shareholders.

(4) Focusing on total debt as opposed to long-term debt is important in this context because the local long-term debt market is small. As a result, personal relationships with suppliers and other short-term creditors can play an important role in the firm's capital structure.

(5) Moreover, although shareholders' decision to adopt a certain legal form of organization is not exogenous, it is unlikely that there is a reverse causation between the choice to incorporate a business and the choice to establish a board. Although firms have to set up a board once they incorporate, if they later switch to a legal form of organization that does not require having a board such as a simplified joint stock company (Sociedades por Acetones Simplificadas, the most common of such switches), few firms dismantle the boards afterward.
Table A1. Survey Questions Used to Construct the Variables

Annual General Meeting (AGM) Index
For the partners' or shareholders' Annual General Meeting during year
20XX, did the company implement the following measures?
1. A document different from the Bylaws, containing the operating rules
of the partners' or shareholders' Annual General Meeting.
2. Calling the meeting with the advance notice stipulated by the Bylaws
and the law.
3. Using other means for the dissemination of the Meeting Call
different from those stipulated by the Bylaws and the Law (email,
website, etc.).
4. The topics to be discussed at the meeting are itemized in the agenda
of the Meeting Call.
5. Including in the Meeting Call the place, time, and manner in which
partners and shareholders may exercise their right of inspection.
6. Appointment of a person delegated by management to ensure the
effectiveness of the right of inspection.
7. Implementing mechanisms for the (annual, biannual) updating of the
data about the partners or shareholders

At the partners' or shareholders' meetings during the year:
8. Is it true that the partners or shareholders are not represented by
managers and employees of the company?
9. Are the year-end financial statements approved by partners who don't
work in the company as managers or employees?
10. Are there documented procedures for the submission and verification
of credentials for the relevant powers?
11. Was the projected distribution of profits presented?
12. Was the proposed distribution of profits approved by a vote equal
to or greater than 78% of the shares or units represented at the
meeting?

Regarding the minutes of the meetings held by partners or shareholders
during the year, has the company established the following measures?
13. Doe the minutes need to be signed by the Chairman and the Company
Secretary?
14. Once approved, are meeting minutes recorded in the appropriate book
within two months of the meeting's celebration?
15. Does the policy for directors' compensation need to be approved by
the Annual Shareholder or Partner Meeting?

Transparency Index
1. Have management reports been discussed at the latest regular
meetings?
2. Do management reports explicitly mention any related party
transactions taken place between the company and its shareholders or
managers?
3. Do management reports include all information about the overall
executive compensation policy, including salaries and other forms of
compensation earned during the year, whatever their cause, in cash or
in kind?

Has the company established any of the following measures with respect
to its partners or shareholders?
4. Documented procedures for responding to requests, questions,
suggestions, or comments made by the partners or shareholders.
5. Mechanisms to publicize clearly, accurately, and thoroughly the
rights and obligations inherent to the condition of partner or
shareholder.
6. Des the company have a website?
7. Are the company's mission and vision statements published on its
website?
8. Are the company bylaws published on its website?
9. Is the company's Good Corporate Governance Code published on its
website?
10. Is the company's Annual Corporate Governance Report published on
its website?
11. Is there a space on the website, where partners or shareholders can
consult financial information containing at least the management
report, the balance sheet, and the income statement for the company's
last fiscal year?
12. Is the tax auditor's opinion about the company's financial
statements disclosed to partners or shareholders more than fifteen days
in advance of the meeting at which they have to approve those
statements?
13. Does the company have as a procedure the rotation of its tax
auditors at least every five years? (When tax auditors are firms, does
the company request that individuals rotate every five years?)
14. When a new tax auditor is to be appointed, is there a minimum of
two proposals of firms or individuals presented to partners or
shareholders at their annual meeting so that they can assess the choice
of the tax auditor?

Top Management Team (TMT) Recruitment & Compensation index
1. Does the company use well-documented and objective selection
processes to hire its key executives?
2. Does the company, through an induction process, provide its newly
appointed legal representatives and key executives with the information
they need to adequately perform their role and the responsibilities
derived from it?
3. Is there a compensation policy for legal representatives and key
executives?
4. Does the compensation policy for legal representatives and key
executives need to be approved by the AGM, partners' meeting, or board
of directors?
5. Is it true that there aren't any forms of compensation, assumption
of expenses, or in-kind benefits other than those approved by the
compensation policy?

Board Independence Index
For the appointment of directors to the board:
1. Are a person's experience, qualifications, and professional prestige
taken into account?
2. Are employees or directors of companies with the same or a similar
business purpose as the company's excluded from consideration?
3. Is it true that the majority required for the board of directors to
make decisions (primary or alternate directors) is not represented by
people linked together by marriage or by blood up to the third degree
of consanguinity (parents, children, grandparents, siblings,
grandchildren, great-grandparents, uncles/aunts, nephews/nieces,
great-grandchildren), second degree of affinity (parents-in-law,
sons/daughters-in-law, siblings-in-law). or first degree of civil
relatedness (adoptive parents, adopted children)?
4. Regarding conflicts of interests: Do directors report any direct or
indirect relationships they may have with any interest group from which
a conflict of interest situation may arise or which may influence the
direction of their opinion or vote?

Board Role Index
Does the board of directors, for the exercise of its functions:
   Board Supervision Subindex
1. Have an operating rulebook?
2. Verify that accounting procedures conform to the legal principles
and that the accounts reflect the actual economic situation of the
company?
   Board Strategic Involvement Subindex
3. Approve a short- and medium-term strategic plan that includes
specific activities and planned implementation dates?
4. Adopt corrective measures designed to guide management towards
meeting the strategic objectives of the company?
5. Submit to the partners' or shareholders' meeting a report describing
the degree of compliance with the strategic objectives of the company?

Board Functioning Index
For board of directors' meetings:
1. Is the information related to directors' responsibilities, duties,
and attributions derived from their office made available to its newly
appointed members?
2. Is there a documented procedure whereby its members receive the
information necessary for making decisions, at least two (2) days in
advance, and according to the meeting agenda?
3. Are there mechanisms that allow alternate members to stay adequately
informed of the issues under consideration by the Board, so that when
replacing the primary members they have the knowledge necessary to
perform their duty?

During the past year:
4. Has an evaluation of the board as a whole been conducted?
5. Has an individual assessment of each board member been conducted?
6. Has the evaluation of the board been performed by an independent
professional or through a Nominating and Compensation Committee?
7. Is it true that none of the primary members of the Board have been
absent from more than three consecutive meetings?
8. Did the board of directors become aware of a measurement of the
performance of the company's legal representatives and executives down
to the second level

Other questions:
9. Does the board have an Audit Committee?
10. Does the board have a Nominating and Compensation Committee?
11. Does the board have a Corporate Governance Committee?
12. Has the company established documented procedures for the
disclosure and management of conflicts of interest?

Table A2. Switches in Legal Form of Organization and Switches in Having
a Board or Not Having a Board
This table reports the number of firm-years in which firms switch from
"without a board" to "with a board" (from "with a board" to "without a
board"). Firms that do not switch are shown in italics on the diagonal.
SA = Sociedad Anonima; SL = Sociedad Limitada; SAS = Sociedadpor
Acciones Simplificada; SU = Sociedad Unipersonal; SCS = Sociedad en
Comandita Simple; SCA = Sociedad en Comandita por Acciones; SC =
Sociedad Colectiva; LFO = legal form of organization.

                                                 Not Legally Required
                                                 to Have a Board:
                                                 With Limited
                                                 Liability
Legal Forms of Organization    With Limited
                               Liability and
                               Legally Required
                               to Have a Board:
                               SA                SL     SAS    SU

With limited liability           0                   0      0    0
and legally required
to have a board:
  SA (corporation)             (0)                 (7)  (603)  (1)
Not legally required to
have a board:
  Companies with limited
  liability, of which:
    SL (limited liability      169                 582    149    0
    corporation)
                               (0)               (975)  (162)  (0)
    SAS (simplified joint        1                   0     70    0
    stock company)
                               (0)                 (1)  (157)  (0)
    SU (single-person            1                   1      0    2
    company)
                               (0)                 (0)    (2)  (4)
  Companies with unlimited
  liability, of which:
    SCS (limited partnership)    5                   1      9    0
                               (0)                 (0)   (11)  (0)
    SCA (limited joint stock     1                   0      4    0
    partnership)
                               (0)                 (0)    (71  (0)
    SC (general partnership)     0                   0      0    0
                               (0)                 (0)    (0)  (0)
Total switches in LFO          177                   2    162    0
(i.e., sum of off-diagonal
cells)
                               (0)                 (8)  (785)  (1)
Total                          177                 584    232    2
                               (0)               (983)  (942)  (5)

                               Not Legally
                               Required to
                               Have a Board:
                               With Unlimited
                               Liability
Legal Forms of Organization

                                                     Total
                                                     Switches in
                               SCS   SCA    SC       LFO          Total

With limited liability            0     0     0        0               0
and legally required
to have a board:
  SA (corporation)              (1)   (0)   (0)    (612)           (612)
Not legally required to
have a board:
  Companies with limited
  liability, of which:
    SL (limited liability         0     0     0      318             900
    corporation)
                                (3)   (0)   (0)    (165)         (1,140)
    SAS (simplified joint         0     0     0        1              71
    stock company)
                                (0)   (2)   (0)      (3)           (160)
    SU (single-person             0     0     0        2               4
    company)
                                (0)   (0)   (0)      (2)             (6)
  Companies with unlimited
  liability, of which:
    SCS (limited partnership)    54     2     0       17              71
                               (75)   (3)   (0)     (14)            (89)
    SCA (limited joint stock      1    28     0        6              34
    partnership)
                                (4)   33)   (0)     (11)            (44)
    SC (general partnership)      0     0     1        0               1
                                (0)   (0)   (0)      (0)             (0)
Total switches in LFO             1     2     0      344
(i.e., sum of off-diagonal
cells)
                                (8)   (5)   (0)    (807)
Total                            55    30     1                    1,081
                               (83)  (38)   (0)                  (2,051)

Table I. Distribution of the Sample across Industries and Legal Forms
of Organization

                                        Number   Percent

Panel A. Industry Classification

 1. Agriculture, hunting and forestry    3,496     6.32%
 2. Fishing                                 63     0.11%
 3. Mining and quarrying                   574     1.04%
 4. Manufacturing                       11,710    21.16%
 5. Electricity, gas and water supply      145     0.26%
 6. Construction                         5,235     9.46%
 7. Commerce (wholesale and retail      17,784    32.14%
    trade, among others)
 8. Hotels and restaurants               1,113     2.01%
 9. Transport, storage and               1,886     3.41%
    communications
10. Investment firms                     2,300     4.16%
    (investment vehicles)
11. Real estate, renting, and            9,559    17.28%
    business activities
12. Education                              296     0.54%
13. Social and health services             119     0.22%
14. Other community, social              1,033     1.87%
    and personal service activities
    Total number of firm-year           55,313   100%
    observations

Panel B. Legal Forms of Organization (US equivalent or translation)

With limited liability and legally
required to have a board:
  Sociedad Anonima (corporation)                23,529    42.54%
Not legally required to have a board:
  Companies with limited liability, of which:
    Sociedad Limitada (limited                  22,267    40.26%
    liability corporation)
    Sociedadpor Acciones                         5,966    10.79%
    Simplificada (simplified
    joint stock company)
    Sociedad Unipersonal                           276     0.50%
    (single-person company)
  Companies with unlimited
  liability, of which:
    Sociedad en Comandita                        2,124     3.84%
    Simple (limited partnership)
    Sociedad en Comandita                        1,136     2.05%
    por Acciones (limited
    joint stock partnership)
    Sociedad Colectiva                              15     0.03%
    (general partnership)
Total number of firm-year                       55,313   100%
observations

Table II. Variable Descriptions

Variable                   Description

Board-related measures
  Corporation Dummy        Dummy variable equal to 1 when the firm is
                           legally organized as a corporation (Sociedad
                           Anonima), and 0 otherwise
  Board Dummy              Dummy variable equal to 1 if the firm has a
                           board of directors, and 0 otherwise
  Board Independence       Percentage of affirmative answers out of the
  Index                    total number of answers given to the 4
                           Colombian Corporate Governance and Social
                           Responsibility Survey (hereafter, Survey)
                           questions related to board independence
  Board Role Index         Percentage of affirmative answers out of the
                           total number of answers given to the 5 Survey
                           questions related to the tasks boards should
                           perform
  Board Supervision        Percentage of affirmative answers out of the
  Subindex                 total number of answers given to the 2 Survey
                           questions related to the monitoring tasks
                           that boards should perform
  Board Strategic          Percentage of affirmative answers out of the
  Involvement              total number of answers given to the 3 Survey
  Subindex                 questions related to the board's involvement
                           in the firm's strategy
  Board Functioning        Percentage of affirmative answers out of the
  Index                    total number of answers given to the 12
                           Survey questions related to functioning of
                           the board

Measures of controlling shareholders' power and identity

  Number of                Total number of shareholders in the firm in
  Shareholders             year t
  Herfindahl Ownership     Sum of the squares of the 5 largest equity
  Concentration Index      stakes in the firm
  Largest Shareholder's    Fraction of cash-flow rights held by the
  Equity Stake             largest shareholder in the firm
  Largest Shareholder      Dummy variable equal to 1 when the largest
  CEO Duality              shareholder is the chief executive officer
                           (CEO), and 0 otherwise
  Largest Shareholder      Dummy variable equal to 1 when the largest
  Director Duality         shareholder is a board member, and 0
                           otherwise
  Family Firm              Dummy variable equal to 1 if the firm states
                           in the Survey that there is financial or
                           managerial family control, and 0 otherwise
  First-Generation         Dummy variable equal to 1 if the firm states
  Family Firm              in the Survey that there is financial or
                           managerial family control exerted by the
                           first generation, and 0 otherwise.
  Second- or               Dummy variable equal to 1 if the firm states
  Later-Generation         in the Survey that there is financial or
  Family Firm              managerial family control exerted by the
                           second or later generation, and 0 otherwise.

Measures of minority shareholders' power

  Annual General           Percentage of affirmative answers out of the
  Meeting (AGM)            total number of answers given to the 15
  Index                    Survey questions related to AGMs
  Nonlargest Shareholder   Dummy variable equal to 1 when one of the 5
  CEO Duality              largest shareholders other than the very
                           largest is the CEO, and 0 otherwise
Variable                   Description
  Nonlargest Shareholder   Dummy variable equal to 1 when one or more of
  Directors Dummy          the 5 largest shareholders other than the
                           very largest serve on the board, and 0
                           otherwise
  Transparency Index       Percentage of affirmative answers out of the
                           total number of answers given to the 14
                           Survey questions related to information
                           disclosure
  Top Management           Percentage of affirmative answers out of the
  Team (TMT)               total number of answers given to the 5 Survey
  Recruitment &            questions related to key managers'
  Compensation             recruitment and compensation policies
  Index
  Foreigners' Equity       Fraction of cash-flow rights held by
  Stake                    foreigners
Dependent variables
  Return on Assets         Net income divided by total assets
  (ROA)
  Dividend Ratio           Dividend paid out by the firm in year t,
                           divided by total assets at the end of year t
  Dividend Dummy           Dummy variable equal to 1 in the year when a
                           firm pays out a dividend, and 0 otherwise
  Leverage                 Ratio of total liabilities divided by total
                           assets
Control variables
  Firm Size                Firm's total assets in US$ million
  Firm Age                 Number of years since the firm's inception
  Growth Opportunities     Percentage growth in real sales
  Volatility of            Standard deviation of the firm's operating
  Operating Profits        margin over the previous three years
  Need for External        Difference between the firm's actual growth
  Funding                  rate and the sustainable growth rate with
                           retained earnings and short-term and
                           long-term debt financing that maintain a
                           constant debt-to-assets ratio (following
                           Durnev and Kim, 2005). We estimate a firm's
                           actual growth rate as a two-year geometric
                           average of annual growth rate in total
                           assets, and the sustainable growth rate as a
                           two-year average of return on equity
                           (ROE)/(1--ROE)

Table III. Descriptive Statistics

This table reports summary statistics of the data. The sample period is
from 2007 to 2012.

                            Means (Standard Deviations)
Variable                    All Firms    Firms         Noncorporations
                                         with Board    with Board
                               (1)         (2)             (3)

Board-related measures
  Corporation Dummy               0.43          0.75         0.00
                                 (0.49)        (0.44)       (0.00)
  Board Dummy                     0.56          1.00         1.00
                                 (0.50)        (0.00)       (0.00)
  Board Independence           --               0.64         0.55
  Index
                               --              (0.27)       (0.27)
  Board Role Index             --               0.71         0.66
                               --              (0.33)       (0.33)
  Board Supervision            --               0.70         0.64
    Subindex                   --              (0.33)       (0.32)
  Board Strategic              --               0.71         0.68
    Involvement Subindex       --              (0.39)       (0.40)
  Board Functioning Index      --               0.37         0.36
                               --              (0.21)       (0.22)
Measures of controlling
shareholders' power
and identity
  Number of Shareholders         10.76         16.50         4.03
                               (132.9)       (177.4)        (3.7)
  Herfindahl Ownership            0.45          0.41         0.48
    Concentration Index          (0.25)        (0.24)       (0.25)
  Largest Shareholder's           0.53          0.50         0.55
    Equity Stake                 (0.24)        (0.24)       (0.24)
  Largest Shareholder CEO         0.37          0.28         0.40
    Duality                      (0.48)        (0.45)       (0.49)
  Largest Shareholder             0.07          0.07         0.07
    Director Duality             (0.25)        (0.25)       (0.25)
  Family Firm                     0.48          0.45         0.54
                                 (0.50)        (0.50)       (0.50)
  First--Generation               0.37          0.33         0.43
  Family
    Firm                         (0.48)        (0.47)       (0.49)
  Second- or Later-               0.10          0.12         0.11
    Generation Family            (0.30)        (0.32)       (0.31)
    Firm
Observations                 55,313        31,221        7,932
Measures of minority
shareholders' power
  AGM Index                       0.58          0.61         0.57
                                 (0.16)        (0.16)       (0.16)
  Nonlargest Shareholder          0.24          0.27         0.20
    CEO Duality                  (0.43)        (0.45)       (0.40)
  Nonlargest Shareholder          0.10          0.10         0.07
    Directors Dummy              (0.31)        (0.31)       (0.26)
  Transparency Index              0.42          0.45         0.41
                                 (0.20)        (0.19)       (0.20)
  TMT Recruitment &               0.49          0.57         0.49
    Compensation Index           (0.36)        (0.36)       (0.37)
  Foreigners' Equity              0.07          0.07         0.06
  Stake                          (0.21)        (0.21)       (0.21)
Dependent variables
  ROA                             0.06          0.06         0.05
                                 (0.24)        (0.26)       (0.22)
  Profit Dummy                    0.84          0.84         0.85
                                 (0.37)        (0.37)       (0.36)
  Dividend Ratio                  0.05          0.04         0.04
                                 (0.30)        (0.31)       (0.18)
  Dividend Dummy                  0.35          0.36         0.30
                                 (0.48)        (0.48)       (0.46)
  Total Leverage                  0.48          0.49         0.49
                                 (0.28)        (0.27)       (0.27)
Control variables
  Firm Size                   9,951        13,866        6,199
                            (56,486)      (70,751)     (33,950)
  Firm Age                       19.39         20.20        19.10
                                (12.87)       (14.07)      (11.91)
  Growth Opportunities            0.02          0.02         0.01
                                 (0.62)        (0.62)       (0.64)
  Volatility of Operating         5.70          0.50         0.27
    Profits                    (744.2)        (32.5)        (8.15)
  Need for External              (0.11)        (0.10)       -0.16
  Funding                        (6.35)        (6.55)       (3.54)
Observations                 55,313        31,221        7,932

                            Means (Standard   Differences in
                            Deviations)       Means[t-stats]
Variable                    Firms
                            without
                             Board
                              (4)             (2)-(4)

Board-related measures
  Corporation Dummy               0.01             0.74 (***)
                                 (0.10)         [260]
  Board Dummy                     0.00          --
                                 (0.00)         --
  Board Independence            --              --
  Index
                                --              --
  Board Role Index              --              --
                                --              --
  Board Supervision             --              --
    Subindex                    --              --
  Board Strategic               --              --
    Involvement Subindex        --              --
  Board Functioning Index       --              --
                                --              --
Measures of controlling
shareholders' power
and identity
  Number of Shareholders          3.47            13.04 (***)
                                 (2.16)          [10.87]
  Herfindahl Ownership            0.51            -0.10 (***)
    Concentration Index          (0.24)         [-45.91]
  Largest Shareholder's           0.57            -0.07 (***)
    Equity Stake                 (0.24)         [-33.82]
  Largest Shareholder CEO         0.48            -0.19 (***)
    Duality                      (0.50)         [-47.84]
  Largest Shareholder           --              --
    Director Duality            --              --
  Family Firm                     0.51            -0.05 (***)
                                 (0.50)         [-12.73]
  First--Generation               0.43            -0.09 (***)
  Family
    Firm                         (0.49)         [-22.28]
  Second- or Later-               0.08             0.04 (***)
    Generation Family            (0.27)          [14.49]
    Firm
Observations                 24,092
Measures of minority
shareholders' power
  AGM Index                       0.55             0.05 (***)
                                 (0.16)          [39.78]
  Nonlargest Shareholder          0.20             0.08 (***)
    CEO Duality                  (0.40)          [20.79]
  Nonlargest Shareholder        --              --
    Directors Dummy             --              --
  Transparency Index              0.37             0.07 (***)
                                 (0.21)          [44.13]
  TMT Recruitment &               0.40             0.17 (***)
    Compensation Index           (0.34)          [55.66]
  Foreigners' Equity              0.06             0.01 (***)
  Stake                          (0.21)           [6.76]
Dependent variables
  ROA                             0.06             0.00
                                 (0.22)          [-0.33]
  Profit Dummy                    0.85            -0.01 (***)
                                 (0.36)          [-3.30]
  Dividend Ratio                  0.05             0.00 (*)
                                 (0.29)          [-1.77]
  Dividend Dummy                  0.34             0.02 (***)
                                 (0.48)           [4.35]
  Total Leverage                  0.47             0.02 (***)
                                 (0.28)           [9.54]
Control variables
  Firm Size                   4,876            8,990 (***)
                            (28,158)             [18.62]
  Firm Age                       18.34             1.86 (***)
                                (11.02)          [16.92]
  Growth Opportunities            0.02             0.00
                                 (0.63)           [0.77]
  Volatility of Operating        12.73           -12.23 (*)
    Profits                  (1,141)          [-1.68]
  Need for External              (0.12)            0.01
  Funding                        (6.08)           [0.23]
Observations                 24,092

                           Differences in Means[t-stats]
Variable
                             (3)-(4)

Board-related measures
  Corporation Dummy           --
                              --
  Board Dummy                 --
                              --
  Board Independence          --
  Index
                              --
  Board Role Index            --
                              --
  Board Supervision           --
    Subindex                  --
  Board Strategic             --
    Involvement Subindex      --
  Board Functioning Index     --
                              --
Measures of controlling
shareholders' power
and identity
  Number of Shareholders         0.56 (***)
                               [16.37]
  Herfindahl Ownership          -0.03 (***)
    Concentration Index        [-9.03]
  Largest Shareholder's        --0.02 (***)
    Equity Stake               [-7.90]
  Largest Shareholder CEO       -0.07 (***)
    Duality                   [-11.38]
  Largest Shareholder         --
    Director Duality          --
  Family Firm                    0.03 (***)
                                [4.30]
  First--Generation              0.00
  Family
    Firm                    [-0-17]
  Second- or Later-              0.03 (***)
    Generation Family           [7.30]
    Firm
Observations
Measures of minority
shareholders' power
  AGM Index                      0.02 (***)
                               [10.22]
  Nonlargest Shareholder         0.01
    CEO Duality                 [1.14]
  Nonlargest Shareholder      --
    Directors Dummy           --
  Transparency Index             0.04 (***)
                               [14.52]
  TMT Recruitment &              0.09 (***)
    Compensation Index         [20.55]
  Foreigners' Equity             0.00
  Stake                         [0.67]
Dependent variables
  ROA                           -0.01 (**)
                               [-1.70]
  Profit Dummy                   0.00
                                [0.47]
  Dividend Ratio                -0.01 (*)
                               [-3.54]
  Dividend Dummy                -0.05 (***)
                               [-7.33]
  Total Leverage                 0.02 (***)
                                [4.73]
Control variables
  Firm Size                  1,322 (***)
                                [4.50]
  Firm Age                       0.75 (***)
                                [5.24]
  Growth Opportunities          -0.01
                               [-0.80]
  Volatility of Operating      -12.45
    Profits                 [-0-85]
  Need for External             -0.04
  Funding                      [-0.46]
Observations

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.
(*) Significant at the 0.10 level.

Table IV. Propensity to Have a Board of Directors

This table reports elected coefficients from random-effects probit
regressions. All regressions include a constant as well as the
following control variables: logs of firm size and age, volatility of
operating profits, growth, leverage, return on assets (ROA), need for
external funding, and industry dummies for each of the industries
listed in Panel A of Table I. Variables are defined in Table II. Robust
standard errors are in parentheses.

                               Noncorporations
Variable                      (1)           (2)

Corporation Dummy             --               --
                              --               --
Number of Shareholders        --                 0.16 (***)
                              --                (0.02)
Herfindahl Ownership           -0.67 (***)       0.13
  Concentration Index          (0.12)           (0.16)
Largest Shareholder            -0.33 (***)      -0.39 (***)
CEO Duality
                               (0.05)           (0.07)
Nonlargest Shareholder         -0.20 (***)      -0.25 (***)
CEO
  Duality                      (0.06)           (0.08)
Family Firm                   --               --
                              --               --
Largest Shareholder           --               --
CEO Duality
   (*) Family Firm            --               --
Nonlargest Shareholder        --               --
CEO
  Duality (*) Family Firm     --               --
First-Generation              --               --
Family Firm
                              --               --
Second- or Later-             --               --
  Generation Family Firm      --               --
AGM Index                       1.87 (***)       2.10 (***)
                               (0.14)           (0.17)
Transparency Index             -0.34 (***)      -0.63 (***)
                               (0.11)           (0.13)
TMT Recruitment &               0.25 (***)       0.29 (***)
  Compensation index           (0.06)           (0.07)
Foreigners' Equity Stake        0.07             0.06
                               (0.15)           (0.17)
Wald test                     390.90 (***)     409.47 (***)
Pseudo-[R.sup.2]                0.29             0.37
Observations               23,240           20,955

                           Noncorporations
Variable                   (3)              (4)

Corporation Dummy             --               --
                              --               --
Number of Shareholders          0.16 (***)       0.16 (***)
                               (0.02)           (0.02)
Herfindahl Ownership            0.18             0.19
  Concentration Index          (0.16)           (0.16)
Largest Shareholder            -0.42 (***)      -0.53 (***)
CEO Duality
                               (0.07)           (0.09)
Nonlargest Shareholder         -0.26 (***)      -0.14
CEO
  Duality                      (0.08)           (0.11)
Family Firm                     0.18 (***)       0.13
                               (0.05)           (0.10)
Largest Shareholder           --                 0.20 (*)
CEO Duality
   (*) Family Firm            --                (0.12)
Nonlargest Shareholder        --                 0.21
CEO
  Duality (*) Family Firm     --                (0.14)
First-Generation              --               --
Family Firm
                              --               --
Second- or Later-             --               --
  Generation Family Firm      --               --
AGM Index                       2.10 (***)       2.11 (***)
                               (0.17)           (0.17)
Transparency Index             -0.64 (***)      -0.64 (***)
                               (0.13)           (0.13)
TMT Recruitment &               0.30 (***)       0.30 (***)
  Compensation index           (0.07)           (0.07)
Foreigners' Equity Stake        0.11             0.08
                               (0.17)           (0.18)
Wald test                     420.63 (***)     430.59 (***)
Pseudo-[R.sup.2]                0.37             0.38
Observations               20,955           20,955

                      Noncorporations   All Firms
Variable              (5)               (6)              (7)

Corporation Dummy        --                  6.36 (***)       6.36 (***)
                         --                 (0.23)           (0.23)
Number of                  0.16 (***)        0.14 (***)       0.14 (***)
Shareholders              (0.02)            (0.01)           (0.01)
Herfindahl Ownership       0.18              0.17             0.17
  Concentration           (0.16)            (0.14)           (0.14)
  Index
Largest Shareholder       -0.42 (***)       -0.50 (***)      -0.37 (***)
CEO Duality
                          (0.07)            (0.08)           (0.06)
Nonlargest                -0.26 (***)       -0.15 (*)        -0.23 (***)
Shareholder CEO
  Duality                 (0.08)            (0.09)           (0.07)
Family Firm              --                  0.08           --
                         --                 (0.09)          --
Largest Shareholder      --                  0.23 (**)      --
CEO Duality
   (*) Family Firm       --                 (0.11)          --
Nonlargest               --                 -0.13           --
Shareholder CEO
  Duality (*) Family     --                 (0.12)          --
  Firm
First-Generation           0.18 (***)      --                 0.15 (***)
Family Firm
                          (0.05)           --                (0.05)
Second- or Later-          0.22 (**)       --                 0.24 (***)
  Generation Family       (0.09)           --                (0.08)
  Firm
AGM Index                  2.10 (***)        2.08 (***)       2.07 (***)
                          (0.17)            (0.15)           (0.15)
Transparency Index        -0.64 (***)       -0.78 (***)      -0.78 (***)
                          (0.13)            (0.11)           (0.11)
TMT Recruitment &          0.30 (***)        0.31 (***)       0.31 (***)
  Compensation index      (0.07)            (0.07)           (0.07)
Foreigners' Equity         0.11              0.03             0.07
Stake                     (0.17)            (0.15)           (0.15)
Wald test                420.80 (***)      993.84 (***)     986.93 (***)
Pseudo-[R.sup.2]           0.37              0.60             0.60
Observations          20,955            37,630           37,630

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.
(*) Significant at the 0.10 level.

Table V. Within-Firm Switches for Having a Board or Not Having a Board

                 Firms with a        Firms without
                 Board That          a Board That
                 Never Switch        Never Switch
Year  Total No.  Number  Percentage  Number  Percentage
      of Firms

2007   5,607      3,413  61%          2,194  39%
2008   4,299      3,018  70%          1,156  27%
2009  16,878      9,326  55%          6,850  41%
2010  16,862      8,921  53%          6,728  40%
2011   5,663      2,720  48%          2,393  42%
2012   6,004      2,742  46%          2,720  45%
All   55,313     30,140  54%         22,041  40%

      Firms That Switch from  Firms That Switch from
      "without a Board" to    "with a Board" to
      "with a Board"          "without a Board"
Year   Number   Percentage     Number   Percentage

2007
2008     66      2%              59      1%
2009    341      2%             361      2%
2010    465      3%             748      4%
2011     84      2%             461      8%
2012    120      2%             422      7%
All   1,081      2%           2,051      4%

Table VI. Propensity to Switch from Having a Board or Not Having a Board

This table reports selected coefficients from random-effects probit
regressions. All regressions include a constant as well as the
following control variables: logs of firm size and age, volatility of
operating profits, growth, leverage, return on assets (ROA), need for
external funding, and industry dummies for each of the industries
listed in Panel A of Table I. Variables are defined in Table II. Robust
standard errors are in parentheses.

                                 Switch to Having
                                 a Board
Variable                         (1)               (2)

Corporation Dummy                     3.00 (***)        3.00 (***)
                                     (0.25)            (0.25)
Number of Shareholders                0.01              0.01
                                     (0.03)            (0.03)
Herfindahl Ownership                  0.08              0.07
Concentration
Index                                (0.10)            (0.10)
Largest Shareholder CEO Duality      -0.02             -0.09
                                     (0.08)            (0.05)
Nonlargest Shareholder CEO            0.05             -0.02
Duality
                                     (0.09)            (0.06)
Family Firm                           0.21 (**)        -
                                     (0.09)            -
Largest Shareholder CEO              -0.14             -
Duality
(*) Family Firm                      (0.11)            -
Nonlargest Shareholder CEO           -0.14             -
Duality
(*) Family Firm
                                     (0.12)            -
First-Generation Family Firm        -                   0.11 (**)
                                    -                  (0.05)
Second- or Later-Generation         -                   0.11
Family Firm                         -                  (0.08)
AGM Index                             0.53 (***)        0.54 (***)
                                     (0.16)            (0.16)
Transparency Index                    0.09              0.09
                                     (0.11)            (0.11)
TMT Recruitment &                    -0.02             -0.03
Compensation
Index                                (0.07)            (0.07)
Foreigners' Equity Stake             -0.41 (***)       -0.44 (***)
                                     (0.15)            (0.15)
Wald test                           398.63            396.73
Pseudo-[R.sup.2]                      0.09              0.09
Observations                     11,996            11,996

                                 Switch to Not Having
                                 a Board
Variable                         (3)              (4)

Corporation Dummy                    -3.45 (***)      -3.44 (***)
                                     (0.16)           (0.16)
Number of Shareholders                0.00             0.00
                                     (0.00)           (0.00)
Herfindahl Ownership                  0.85 (***)       0.84 (***)
Concentration
Index                                (0.09)           (0.09)
Largest Shareholder CEO Duality       0.53 (***)       0.42 (***)
                                     (0.07)           (0.05)
Nonlargest Shareholder CEO            0.14 (**)        0.07
Duality
                                     (0.07)           (0.05)
Family Firm                           0.20 (**)       -
                                     (0.08)          -
Largest Shareholder CEO              -0.22 (**)      -
Duality
(*) Family Firm                      (0.10)          -
Nonlargest Shareholder CEO           -0.18           -
Duality
(*) Family Firm
                                     (0.10)          -
First-Generation Family Firm         -                 0.07
                                     -                (0.04)
Second- or Later-Generation          -                -0.05
Family Firm                          -                (0.07)
AGM Index                            -1.24 (***)      -1.23 (***)
                                     (0.15)           (0.15)
Transparency Index                    0.21 (**)        0.21 (**)
                                     (0.11)           (0.11)
TMT Recruitment &                    -0.22 (***)      -0.22 (***)
Compensation
Index                                (0.06)           (0.06)
Foreigners' Equity Stake             -0.25 (**)       -0.29 (**)
                                     (0.12)           (0.12)
Wald test                         2,249.30         2,246.85
Pseudo-[R.sup.2]                      0.30             0.30
Observations                     11,216           11,216

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.

Table VII. Effects of Having a Board on Financial Policies and
Performance

This table reports selected regression coefficients. The last three
columns report second-stage estimates from treatment effects regression
models, where the first stage follows the specification in Column 6 of
Table IV. All regressions include a constant as well as the following
control variables: logs of firm size and age, volatility of operating
profits, growth, leverage, return on assets (ROA), need for external
funding, and industry dummies for each of the industries listed on
Panel A of Table I. Variables are defined in Table II. Robust standard
errors are in parentheses, and p-values are in brackets.

Variable                       Random            Tobit
                               Effects           Dividend
                               Leverage          Ratio

Board Dummy                         0.01 (***)        -0.06 (***)
                                   (0.00)             (0.01)
Herfindahl Ownership                0.03 (***)        -0.04 (***)
Concentration
Index                              (0.01)             (0.01)
Largest Shareholder CEO             0.01 (**)          0.03 (***)
Duality
                                   (0.00)             (0.01)
Largest Shareholder Director        0.01 (*)          -0.02
Duality
                                   (0.01)             (0.02)
Nonlargest Shareholder CEO          0.00               0.03 (***)
Duality
                                   (0.00)             (0.01)
Nonlargest Shareholder              0.00              -0.04 (**)
Directors
Dummy                              (0.00)             (0.02)
Family Firm                         0.01              -0.00
                                   (0.00)             (0.01)
Largest Shareholder CEO            -0.01 (*)           0.02 (*)
Duality
(*) Family Firm                    (0.01)             (0.01)
Nonlargest Shareholder             -0.01 (**)          0.01
CEO
Duality (*) Family Firm            (0.01)             (0.01)
AGM Index                          -0.04 (***)         0.57 (***)
                                   (0.01)             (0.02)
Transparency Index                  0.05 (***)         0.18 (***)
                                   (0.01)             (0.01)
TMT Recruitment &                   0.01 (***)        -0.04 (***)
Compensation
Index                              (0.00)             (0.01)
Foreigners' Equity Stake           -0.04 (***)         0.01
                                   (0.01)             (0.01)
Pseudo-[R.sup.2]                    0.24               0.27
Wald [chi-square] test          3,990            [12,617]
Lambda                         -                 -
                               -                 -
Observations                   41,831             41,831

                                                2nd-Stage Treatment
                                                Effects
Variable                       Random           Leverage
                               Effects
                               ROA

Board Dummy                         0.01 (***)       0.02 (***)
                                   (0.00)           (0.00)
Herfindahl Ownership                0.01             0.03 (***)
Concentration
Index                              (0.01)           (0.01)
Largest Shareholder CEO            -0.01 (*)         0.04 (***)
Duality
                                   (0.01)           (0.00)
Largest Shareholder Director        0.01 (**)        0.02 (*)
Duality
                                   (0.01)           (0.01)
Nonlargest Shareholder CEO         -0.01             0.03 (***)
Duality
                                   (0.01)           (0.00)
Nonlargest Shareholder             -0.01 (**)        0.02 (*)
Directors
Dummy                              (0.01)           (0.01)
Family Firm                        -0.01             0.01
                                   (0.01)           (0.01)
Largest Shareholder CEO             0.01            -0.02 (***)
Duality
(*) Family Firm                    (0.01)           (0.01)
Nonlargest Shareholder              0.01            -0.02 (***)
CEO
Duality (*) Family Firm            (0.01)           (0.01)
AGM Index                           0.08 (***)      -0.11 (***)
                                   (0.01)           (0.01)
Transparency Index                  0.15 (***)       0.06 (***)
                                   (0.01)           (0.01)
TMT Recruitment &                   0.01 (**)        0.02 (***)
Compensation
Index                              (0.00)           (0.00)
Foreigners' Equity Stake            0.04 (***)      -0.03 (***)
                                   (0.01)           (0.01)
Pseudo-[R.sup.2]                    0.06        -
Wald [chi-square] test          1,010           11,717
Lambda                         -                    -0.01 (***)
                               -                    (0.00)
Observations                   41,831           35,681

                               2nd-Stage Treatment
                               Effects
Variable                       Dividend         ROA
                               Dummy

Board Dummy                        -0.09 (***)       0.02 (***)
                                   (0.01)           (0.00)
Herfindahl Ownership               -0.08 (***)       0.01
Concentration
Index                              (0.01)           (0.01)
Largest Shareholder CEO             0.04 (***)      -0.02 (***)
Duality
                                   (0.01)           (0.01)
Largest Shareholder Director       -0.03             0.01
Duality
                                   (0.02)           (0.01)
Nonlargest Shareholder CEO          0.04 (***)      -0.02 (***)
Duality
                                   (0.01)           (0.01)
Nonlargest Shareholder             -0.06 (***)      -0.01
Directors
Dummy                              (0.02)           (0.01)
Family Firm                         0.04 (***)       0.00 (*)
                                   (0.01)           (0.01)
Largest Shareholder CEO            -0.02            -0.00
Duality
(*) Family Firm                    (0.01)           (0.01)
Nonlargest Shareholder             -0.02 (*)        -0.00
CEO
Duality (*) Family Firm            (0.01)           (0.01)
AGM Index                           0.88 (***)       0.08 (***)
                                   (0.02)           (0.01)
Transparency Index                  0.24 (***)       0.15 (***)
                                   (0.01)           (0.01)
TMT Recruitment &                  -0.07 (***)       0.01 (*)
Compensation
Index                              (0.01)           (0.01)
Foreigners' Equity Stake           -0.07 (***)       0.04 (***)
                                   (0.01)           (0.01)
Pseudo-[R.sup.2]               -                -
Wald [chi-square] test          8,657            2,977
Lambda                             -0.00            -0.01 (***)
                                   (0.01)           (0.00)
Observations                   35,681           35,681

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.
(*) Significant at the 0.10 level.

Table VIII. Treatment Effects of Switching to Having a Board or Not
Switching to Having a Board on Financial Policies and Performance

This table reports selected second-stage coefficients from treatment
effects regression models. The first stage follows the specification in
Column 1 of Table VI (for the switches to having a board--first three
columns) or Column 3 (for the switches to not having a board--last
three columns). All regressions include a constant as well as the
following control variables: logs of firm size and age, volatility of
operating profits, growth, leverage, return on assets (ROA), need for
external funding, and industry dummies for each of the industries
listed on Panel A of Table I. Variables are defined in Table II. Robust
standard errors are in parentheses, and p-values are in brackets.

Variable                     Switch to Having a Board
                             Leverage         Dividend
                                              Dummy

Board Switch Dummy                0.17 (***)      -0.15 (*)
                                 (0.04)           (0.07)
Herfindahl Ownership              0.02 (*)        -0.07 (***)
Concentration Index              (0.01)           (0.02)
Largest Shareholder CEO           0.02 (***)       0.06 (***)
Duality                          (0.01)           (0.02)
Non-Largest Shareholder           0.01             0.03 (*)
CEO Duality                      (0.01)           (0.02)
Family Firm                      -0.01             0.06 (***)
                                 (0.01)           (0.02)
Largest Shareholder CEO           0.01            -0.03
Duality (*) Family Firm          (0.01)           (0.02)
Non-Largest Shareholder          -0.01            -0.02
CEO Duality (*) Family Firm      (0.01)           (0.03)
AGM Index                        -0.11 (***)       0.80 (***)
                                 (0.02)           (0.03)
Transparency Index                0.04 (***)       0.28 (***)
                                 (0.01)           (0.02)
TMT Recruitment &                 0.03 (***)      -0.05 (***)
Compensation Index               (0.01)           (0.01)
Foreigners' Equity Stake         -0.05 (***)      -0.08 (***)
                                 (0.01)           (0.02)
Wald test                     3,382            2,543
Lambda                           -0.07 (***)       0.03
                                 (0.02)           (0.03)
Observations                 10,918           10,918

Variable                     Switch to        Switch to Not Having a
                             Having a Board   Board
                             ROA              Leverage

Board Switch Dummy               -0.31 (***)      -0.04 (***)
                                 (0.04)           (0.02)
Herfindahl Ownership              0.02             0.02 (**)
Concentration Index              (0.01)           (0.01)
Largest Shareholder CEO          -0.02             0.05 (***)
Duality                          (0.01)           (0.01)
Non-Largest Shareholder          -0.02             0.04 (***)
CEO Duality                      (0.01)           (0.01)
Family Firm                       0.02 (*)         0.01
                                 (0.01)           (0.01)
Largest Shareholder CEO          -0.02            -0.04 (***)
Duality (*) Family Firm          (0.01)           (0.01)
Non-Largest Shareholder          -0.01            -0.03 (***)
CEO Duality (*) Family Firm      (0.02)           (0.01)
AGM Index                         0.10 (***)      -0.11 (***)
                                 (0.02)           (0.02)
Transparency Index                0.16 (***)       0.05 (***)
                                 (0.02)           (0.01)
TMT Recruitment &                 0.01             0.03 (***)
Compensation Index               (0.01)           (0.01)
Foreigners' Equity Stake          0.05 (***)      -0.03 (**)
                                 (0.02)           (0.01)
Wald test                       943            4,183
Lambda                            0.14 (***)       0.02 (***)
                                 (0.02)           (0.01)
Observations                 10,918           11,170

Variable                     Switch to Not Having a Board

                             Dividend         ROA
                             Dummy

Board Switch Dummy                0.01            -0.12 (***)
                                 (0.00)           (0.02)
Herfindahl Ownership             -0.09 (***)      -0.00
Concentration Index              (0.02)           (0.02)
Largest Shareholder CEO           0.04 (***)      -0.04 (***)
Duality                          (0.02)           (0.01)
Non-Largest Shareholder           0.05 (***)      -0.03 (***)
CEO Duality                      (0.01)           (0.01)
Family Firm                       0.03 (**)        0.00
                                 (0.02)           (0.01)
Largest Shareholder CEO          -0.02             0.02
Duality (*) Family Firm          (0.02)           (0.02)
Non-Largest Shareholder          -0.02            -0.01
CEO Duality (*) Family Firm      (0.02)           (0.02)
AGM Index                         0.96 (***)       0.12 (***)
                                 (0.03)           (0.03)
Transparency Index                0.25 (***)       0.18 (***)
                                 (0.03)           (0.02)
TMT Recruitment &                -0.08 (***)       0.02
Compensation Index               (0.01)           (0.01)
Foreigners' Equity Stake         -0.07 (***)       0.07 (***)
                                 (0.02)           (0.02)
Wald test                     3,291            1,860
Lambda                            0.04 (***)       0.06 (***)
                                 (0.02)           (0.01)
Observations                 11,170           11,170

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.
(*) Significant at the 0.10 level.

Table IX. Effects of Board's Roles and Effective Functioning on
Financial Policies and Performance

This table reports selected regression coefficients from the second
stage of sample selection models, where the first-stage probit model
follows the specification in Column 6 of Table III. All regressions
include a constant as well as the following control variables: logs of
firm size and age, volatility of operating profits, growth, leverage,
return on assets (ROA), need for external funding, and industry dummies
for each of the industries listed on Panel A of Table I. Variables are
defined in Table II. The number of observations in all regressions is
35,681, of which 14,589 are censored and 21,092 are uncensored. Robust
standard errors are in parentheses.

Variable                     Leverage

Board Independence Index        -0.00           -0.00
                                (0.01)          (0.01)
Board Role Index                 0.02 (**)      -
                                (0.01)          -
Board Supervision Subindex      -               -0.01
                                -               (0.01)
Board Strategic Involvement     -                0.02 (***)
  Subindex                      -               (0.01)
Board Functioning Index          0.02            0.01
                                (0.01)          (0.01)
Herfindahl Ownership             0.03 (***)      0.03 (***)
  Concentration Index           (0.01)          (0.01)
Largest Shareholder              0.06 (***)      0.06 (***)
CEO Duality
                                (0.01)          (0.01)
Largest Shareholder              0.02 (**)       0.02 (**)
Director
  Duality                       (0.01)          (0.01)
Nonlargest Shareholder           0.03 (***)      0.03 (***)
CEO
  Duality                       (0.01)          (0.01)
Nonlargest Shareholder           0.01            0.01
Director
  Duality                       (0.01)          (0.01)
Family Firm                      0.02 (***)      0.02 (***)
                                (0.01)          (0.01)
Largest Shareholder             -0.04 (***)     -0.04 (***)
CEO Duality
  (*) Family Firm               (0.01)          (0.01)
Nonlargest Shareholder          -0.03 (***)     -0.03 (***)
CEO
  Duality (*) Family Firm       (0.01)          (0.01)
ACM Index                       -0.11 (***)     -0.11 (***)
                                (0.01)          (0.01)
Transparency Index               0.06 (***)      0.06 (***)
                                (0.01)          (0.01)
TMT Recruitment &                0.03 (***)      0.03 (***)
Compensation
  Index                         (0.01)          (0.01)
Foreigners' Equity              -0.02 (**)      -0.02 (**)
Stake
                                (0.01)          (0.01)
Wald test                    6,677 (***)     6,688 (***)

Variable                     Dividend Dummy

Board Independence Index        -0.12 (**)      -0.11 (**)
                                (0.05)          (0.05)
Board Role Index                -0.69 (***)     -
                                (0.04)          -
Board Supervision Subindex      -               -0.25 (***)
                                -               (0.04)
Board Strategic Involvement     -               -0.43 (***)
  Subindex                      -               (0.04)
Board Functioning Index         -0.72 (***)     -0.72 (***)
                                (0.07)          (0.07)
Herfindahl Ownership            -0.33 (***)     -0.33 (***)
  Concentration Index           (0.04)          (0.04)
Largest Shareholder              0.09 (**)       0.09 (**)
CEO Duality
                                (0.04)          (0.04)
Largest Shareholder             -0.05           -0.05
Director
  Duality                       (0.06)          (0.06)
Nonlargest Shareholder           0.14 (***)      0.14 (***)
CEO
  Duality                       (0.03)          (0.03)
Nonlargest Shareholder          -0.10 (**)      -0.10 (**)
Director
  Duality                       (0.05)          (0.05)
Family Firm                      0.07 (**)       0.07 (**)
                                (0.04)          (0.04)
Largest Shareholder             -0.05           -0.05
CEO Duality
  (*) Family Firm               (0.05)          (0.05)
Nonlargest Shareholder          -0.12 (**)      -0.12 (**)
CEO
  Duality (*) Family Firm       (0.05)          (0.05)
ACM Index                        3.26 (***)      3.26 (***)
                                (0.08)          (0.08)
Transparency Index               0.68 (***)      0.68 (***)
                                (0.06)          (0.06)
TMT Recruitment &                0.02            0.02
Compensation
  Index                         (0.03)          (0.03)
Foreigners' Equity              -0.14 (***)     -0.14 (***)
Stake
                                (0.05)          (0.05)
Wald test                    4,604 (***)     4,618 (***)

Variable                     ROA

Board Independence Index         -0.00            -0.00
                                 (0.01)           (0.01)
Board Role Index                 -0.05 (***)      -
                                 (0.01)           -
Board Supervision Subindex       -                -0.02 (**)
                                 -                (0.01)
Board Strategic Involvement      -                -0.04 (***)
  Subindex                       -                (0.01)
Board Functioning Index          -0.02 (**)       -0.02 (**)
                                 (0.01)           (0.01)
Herfindahl Ownership              0.00             0.00
  Concentration Index            (0.01)           (0.01)
Largest Shareholder              -0.01            -0.01
CEO Duality
                                 (0.01)           (0.01)
Largest Shareholder               0.02             0.02
Director
  Duality                        (0.01)           (0.01)
Nonlargest Shareholder           -0.00            -0.00
CEO
  Duality                        (0.01)           (0.01)
Nonlargest Shareholder           -0.01            -0.01
Director
  Duality                        (0.01)           (0.01)
Family Firm                       0.01             0.01
                                 (0.01)           (0.01)
Largest Shareholder               0.00             0.00
CEO Duality
  (*) Family Firm                (0.01)           (0.01)
Nonlargest Shareholder           -0.02 (**)       -0.02 (**)
CEO
  Duality (*) Family Firm        (0.01)           (0.01)
ACM Index                         0.12 (***)       0.12 (***)
                                 (0.01)           (0.01)
Transparency Index                0.07 (***)       0.07 (***)
                                 (0.01)           (0.01)
TMT Recruitment &                 0.02 (***)       0.02 (***)
Compensation
  Index                          (0.01)           (0.01)
Foreigners' Equity                0.02 (*)         0.02 (*)
Stake
                                 (0.01)           (0.01)
Wald test                    11,042 (***)     11,044 (***)

(***) Significant at the 0.01 level.
(**) Significant at the 0.05 level.
(*) Significant at the 0.10 level.
COPYRIGHT 2019 Financial Management Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Villalonga, Belen; Trujillo, Maria-Andrea; Guzman, Alexander; Caceres, Neila
Publication:Financial Management
Article Type:Survey
Geographic Code:0LATI
Date:Jun 22, 2019
Words:21100
Previous Article:Expected versus Ex Post Profitability in the Cross-Section of Industry Returns.
Next Article:Do Business Borrowers Benefit from Bank Bailouts?: The Effects of TARP on Loan Contract Terms.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters