Introduction

The separation of ownership and control in modern corporations affords managers wide latitude over use of free cash flow. Self-serving managers prefer to use excess cash to boost personal benefits and compensation, grow through acquisitions, and deter hostile takeovers.Footnote 1 To prevent wealth expropriation by managers, and investment in takeover-resistant, value-destroying projects, various governance mechanisms have been developed. External monitoring is provided by institutional investors and blockholders, and the threat of removal of inefficient management by the market for corporate control. Internal controls include monitoring by an independent board, oversight by board committees, equity-based compensation for managers, and capital structure and payout policies. Dividends mitigate agency costs by reducing free cash flow, and subjecting firms to the discipline by external agents at the time of raising capital. As such, entrenched managers prefer not to pay dividends. We explore how CEO entrenchment, measured by CEO’s power and influence over selection of directors, impacts dividend policy. Specifically, we focus on the hypothesis that entrenched CEOs who feel secure in their jobs pay high dividend as they need not stockpile cash to defend against hostile takeovers.

The premise that dividend payment is effective in mitigating agency costs is well articulated by Easterbrook (1984), and later analyzed by Jensen (1986). Easterbrook (1984) asserts that use of cash for dividend payment forces the firm to raise investment capital from the market, exposing it to the scrutiny of investment bankers and institutional investors who act as monitors for the collective interest of shareholders. Jensen (1986) notes that dividend distribution reduces free cash flow at managers’ disposal, prevents wasteful investments and alleviates agency costs. These models imply that entrenched managers prefer not to pay dividends to minimize disciplining by institutional investors and blockholders. Instead, they conserve funds for private benefits and to repurchase stock and make acquisitions to thwart unwanted takeovers. Therefore, firms with weak (strong) shareholder protection pay low (high) dividends. Empirical evidence contradicts this prediction, however. Hu and Kumar (2004) report significantly positive impact of CEO power on dividend payout. Ghosh and Sirmans (2006) find similar results for REITs. John and Knyazeva (2006) conclude that in firms with weak corporate governance, dividends represent a pre-commitment not to waste cash on non-value-maximizing initiatives. Pan (2006) and Harford et al. (2006) also find that firms with strong managerial power pay higher dividends.

A straightforward explanation for this apparently counter-intuitive result is that dividends and governance mechanisms are substitutes so that shareholders demand high dividends of firms with weak shareholder rights (Easterbrook 1984, and Jensen 1986). Consistent with the substitution hypothesis, Hu and Kumar (2004) argue that entrenched managers pay dividends to avoid shareholder sanctions—shareholders are more tolerant of weak governance when managers have limited cash at their disposal. Similarly, John and Knyazeva (2006) note that dividends and governance controls are at least partial substitutes. Ghosh and Sirmans (2006) come to similar conclusions. However, the substitution hypothesis begs the question: what is the incentive for entrenched managers to pay higher dividends when they can use the funds to enhance their personal wealth, or make acquisitions and/or stock repurchase to fend off hostile suitors? John and Knyazeva (2006) and Harford et al. (2006) conjecture that due to the strict protection and enforcement of shareholder rights in the USA, agency costs associated with cash reserves make entrenched, poorly monitored managers vulnerable to disciplining by the takeover market. So entrenched managers dissipate cash on dividends and takeover-resistant acquisitions even if they are non-value-maximizing.Footnote 2 This explanation, while intuitive, does not describe how shareholders who prefer dividends, induce managers to comply with their preferred option. Any incentive shareholders offer must provide managers adequate job security. Pan (2006) develops the optimal entrenchment hypothesis under which high dividend payment by entrenched managers is an equilibrium solution where shareholders allow managers some level of entrenchment through anti-takeover protections and weak governance as an inducement to use excess cash for dividends rather than on value-destroying activities.Footnote 3 When managers are protected from hostile takeover threats, they feel secure to focus on long-term initiatives instead of myopic short-term goals (i.e. value-destroying investments). The potential gains from such refocusing attenuate the agency cost of managerial entrenchment.

We study the relationship between CEO entrenchment and dividend policy of real estate investment trusts (REITs).Footnote 4 The impact of CEO entrenchment on dividend payout is particularly relevant for REITs due to the unique regulatory provisions which, as several authors have observed, may have the unintended effect of insulating incumbent management from the disciplinary forces of the takeover market. For instance, the requirement that the five largest owners of a REIT may not own more than 50% of the total outstanding shares impedes blockholdings, and makes takeover threats less likely. The rule that 75% of REIT investments must be held in real estate assets curtails CEO’s experience in diverse industries and his employment potential. The limited opportunity for career change may lead to collusion among CEOs to pre-empt hostile takeovers, fostering entrenchment. As such, the regulatory environment of REITs has the potential to render ineffective the monitoring imposed by external forces. The failure of the external control mechanism to address agency-related issues vests added responsibility on the board of directors to guide management toward optimal behavior or otherwise replace the incumbent team. In situations such as this, CEO’s power and influence over the board can have significant impact on decision making.

In an environment where external monitoring and governance is less effective, do shareholders allow managers greater entrenchment through weak internal governance to maximize dividends? If so, then we would expect dividends and managerial influence over internal control mechanisms to be positively related. To explore this relationship, we focus on the independence of the board of directors. An independent board is critical to close monitoring of managers. Instead of using the standard proxy for board independence—the number of independent directors, we measure board independence in terms of CEO’s influence on director nomination and selection. If the CEO has influence over director selection, outside directors whose reputation and human capital derive from length of service, may be forced to compromise their independence to get reappointed. Outside directors that must remain loyal to the CEO to extend their tenure ultimately lose independence and fail to provide adequate monitoring. Since a captive board is not expected to align with dissidents in a fight for control, takeovers are less likely. Under the reduced threat of takeovers, entrenched managers need not conserve cash to defend against hostile bidders so that they are willing to pay high dividends.

Our analyses of dividend policy of REITs for 1999 and 2000 confirm that entrenched CEOs with greater influence on internal governance pay higher dividends. Our main finding is that the positive relation between CEO entrenchment and dividend policy is driven by REITs with no nomination committee. In the absence of nomination committee, CEO is in complete control of the slate of nominees and likely to nominate candidates that are loyal to him. The analysis reveals that REITs with entrenched CEOs (long tenure and dual role as chairman) and no nomination committee pay the highest dividend. We conclude that even when external control systems are weak, shareholders surrender considerable power to the CEO to design a friendly board as added protection against hostile takeovers to induce a high dividend payment.

This study makes important contributions. The notion that some level of CEO entrenchment is value-maximizing is a relatively new concept. We observed earlier that the regulatory environment of REITs can blunt the threat of hostile takeovers and disciplining by external agencies. Although REITs must pay high dividends by regulation, large depreciation expenses leave considerable cash flow at managers’ discretion. In conjunction, these factors imply that internal monitoring by the board exerts significant influence on managerial decision making. Focusing on internal governance mechanisms, we find that more entrenched CEOs (greater influence on the board) pay higher dividends. In essence, an effective way to induce CEOs to return cash to shareholders is to protect them from hostile threats. A protected CEO is less prone to stockpile cash for personal use or create barriers to hostile threats. Our first contribution is to corroborate the notion that deployment of corporate assets by US managers is guided by the overall environment of better protection and enforcement of shareholder rights, not governance at the firm or industry level (John and Knyazeva 2006 and Harford et al. 2006). Our analysis of REITs demonstrates that the effect prevails even when external governance is weak. Second, we contend that shareholders’ willingness to induce higher dividend distribution by allowing managers additional protection with a weak board can be interpreted as evidence for the optimal entrenchment hypothesis. Finally, our analysis demonstrates the importance of the composition of nomination committee as a determinant of board independence. The data suggest that influence on the slate of nominees gives CEO significant bargaining power over the board. Indeed, once the composition of the nomination committee is accounted for, other sources of CEO power—tenure and duality—are redundant. To our knowledge, this is a new result.

The paper is organized as follows. In the next section, we discuss the extant literature on the governance of REITs and develop our hypotheses. Next, we present the data and results from univariate analysis. The relative importance of CEO character and governance devices for dividend policy is explored after that. Finally, the paper is summarized in the last section.

Background Literature and Hypotheses

Dividend Policy as a Monitoring Device

The divergence of interest between shareholders and managers involves the deployment of free cash flow. The source of the conflict is managerial preference to use excess cash for personal benefits and takeover resistance. Harford et al. (2006) show that firms with weak shareholder rights (i.e. entrenched managers) hold less cash because they dissipate excess cash on acquisitions to thwart hostile takeover attempts. Dittmar and Mahrt-Smith (2006) report that the incremental value of a dollar is significantly less in firms with entrenched managers, reflecting shareholders’ concern that managers waste cash on value-decreasing investments.

Pre-commitment to disburse cash through dividends is of significant benefit to shareholders as a monitoring tool. The pre-commitment is credible because managers are reluctant to cut dividends because of the strong negative reaction. Indeed, dividend payment is the most direct and visible means for managers to disgorge free cash flow. In addition to returning cash to shareholders, dividends force managers to raise capital from external sources and face monitoring by underwriters, investment bankers, and rating agencies at issuance (Easterbrook 1984). By reducing firms’ free cash flow, dividends limit opportunities for wasteful investments. Further, trips to the market make managers’ risk-aversion less costly for shareholders. Entrenched, risk-averse managers avoid risky investments to minimize the potential loss of personal wealth tied up in the firm’s stock. Such a strategy causes transfer of wealth from shareholders to bondholders. Dividend payment compels the firm to raise money and the resulting debt–equity adjustments benefit shareholders.Footnote 5

Nevertheless, despite the monitoring benefits of dividends, self-interested managers prefer to use excess cash to enhance their private benefits, and take actions that make acquisitions more costly for hostile bidders. Faleye (2004) discusses various anti-takeover defense options available to cash rich firms. Stulz (1988) argues that target management can deter takeovers by increasing their control of voting rights through share acquisitions. Bagwell (1991, 1992) demonstrates that the supply curve of shares is upward-sloping such that share repurchases increase the cost of a takeover. Other defensive actions that cash facilitates include acquiring a competitor of the hostile bidder, or the bidder itself. Consistent with this view, Harford (1999) and Pinkowitz (2002) provide evidence that cash holdings deter takeover bids. Assuming that entrenched managers prefer to conserve cash as a defensive mechanism rather than pay dividends, managerial entrenchment and dividend payments should be negatively related.

Interestingly, the evidence from US firms contradicts the notion that firms with entrenched managers (weak shareholder rights) pay lower dividends. Hu and Kumar (2004) and Ghosh and Sirmans (2006) report that more powerful (weakly governed) CEOs pay higher dividends. John and Knyazeva (2006) find that weak external and internal corporate governance is associated with a greater emphasis on dividends. Officer (2006) finds that predicted dividend payers with weak governance are significantly more likely to initiate dividend than are predicted dividend payers with strong governance. The authors assert that cash distribution is not as critical to contain deadweight agency costs when efficient governance limits managers’ opportunity to expropriate corporate resources. Conversely, when agency costs of free cash flow are high due to strong managerial rights, shareholders prefer the pre-commitment feature of dividends as a restriction on deviant managerial actions. Under this interpretation, dividends and governance mechanisms are substitutes in mitigating agency costs.

The substitution hypothesis is appealing, but two issues merit attention. First, unlike the US evidence, international evidence on the impact of governance on dividend policy is consistent with the prediction that poorly (strongly) governed managers pay lower (higher) dividends (La Porta et al. (2000), and Dewenter and Warther (1998)). Why do US managers of firms with weak shareholder rights pay high dividends? Harford et al. (2006) and John and Knyazeva (2006) attribute the disparate findings to the strong governance system in the USA. According to them, strong protection and enforcement of investors’ rights in the USA, and vulnerability of managers to a performance-based turnover system force even poorly monitored managers to maintain dividends at a pre-committed level. The pre-commitment is credible because dividend cuts signal poor performance that hostile acquirers can capitalize on. Similarly, managers are unlikely to increase dividends until they are confident that the earnings level is sustainable. Consequently superior firms will accumulate cash if they do not pay dividends. Because of the associated agency costs, large cash stockpiles can trigger shareholder sanctions.Footnote 6 To quell shareholder outrage over accumulation of cash, entrenched managers dissipate cash on non-value-maximizing investments (Harford et al. 2006), or make dividend payment (Hu and Kumar 2004). To continue enjoying the private benefits of power and prestige, entrenched managers are the most anxious not to incur shareholder wrath over excessive accumulation of cash. As such, entrenched managers unload cash stockpile quickly. Conversely, absent the active market for corporate control in international markets, managers face less pressure to pay dividends. Harford et al. (2006) argue that what matters is the intensity of monitoring by the capital market, not the strength of shareholder rights at the firm-level. Similarly, Doidge et al. (2004) observe that mechanisms for a firm to commit to higher quality governance may be unavailable or prohibitively expensive in countries with poor investor protection and poor economic development. So, if the tools required to implement good governance practices are not available in the country, firms have no incentive to impose controls on managerial behavior. Given that the main benefit of good corporate governance is easier access to capital market, there is no motivation for corporate governance in a country where the capital market is not yet developed. In essence, in a country where financial development and investor protection are poor, it is not worth for individual firms to bond themselves to superior governance.

Although both dividends and acquisitions dissipate cash, a self-interested manager prefers acquisitions because in addition to making hostile takeovers less likely, growth enhances managerial benefits. As such, shareholders must provide appropriate incentives to induce managers to payout excess cash as dividends instead of investing it in value-destroying projects. Pan (2006) presents a model where shareholders’ surrender sufficient power to managers through anti-takeover provisions and/or weak internal governance so that managers feel secure and choose to use cash stockpile to pay dividends instead of making value-destroying, takeover-resistant acquisitions. Pan (2006) develops the optimal entrenchment hypothesis which states that there is an optimal, value-maximizing level of entrenchment where cost of dividends is offset by the gains from managers’ commitment not to undertake value-destroying, myopic investments.Footnote 7 Similar in spirit to Pan (2006), Almazan and Suarez (2003) present a model where the combination of weak board and large severance pay maximizes shareholder wealth. A weak board allows the CEO to be entrenched. However, the diligent CEO knows that the probability of his reappointment and opportunity to renegotiate a bigger severance pay when a modestly better candidate emerges is higher with superior performance. A negligent and incompetent CEO, on the other hand, is aware of his motives and capabilities and will accept the severance package and leave. The authors show that such an arrangement is ex-ante shareholder-value-maximizing.

Regulatory Structure and Governance of REITs

We now explore the implications of the special regulatory provisions of REITs for the two hypotheses described above. The severity of agency problem depends on the information asymmetry between shareholders and managers, the efficacy of governance mechanisms, and the threat of takeover by large shareholders and hostile acquirers. It is often argued that because management has less discretion in a regulated industry, information asymmetry is less severe. There is no consensus on information asymmetry in the REIT sector, however.Footnote 8 Some authors argue that heterogeneity and illiquidity of real estate assets make valuation difficult, while others view the value of a REIT as simply the aggregate value of its assets. The divergent opinions notwithstanding, most agree that the increase in institutional holding over the last few years has improved the transparency of REITs.

Agency problem is generally less severe in regulated industries because regulation diminishes managerial discretion. However, regulation of REITs originated not out of concern for fiduciary responsibility and protection of stakeholder interest, but to improve liquidity, diversification, and transparency to attract institutional investors. Consider the regulatory provisions of REITS with potential impact on governance:

  1. 1.

    Ninety percent distribution rule: For tax exemption, 90% of taxable earnings must be paid out as dividends. This rule reduces managers’ access to discretionary cash flow, and forces firms to raise capital from the market, exposing them to the disciplining forces of investment bankers. Consequently, agency conflict is mitigated.

  2. 2.

    Restriction on income sources: 75% of assets must be held in cash, government securities, and real estate assets. Restricting diversification to the real estate sector limits managers’ experience and employment potential. To protect their careers, REIT managers likely collude to thwart hostile takeovers (Campbell et al. 2001). The resulting entrenchment increases agency costs.

  3. 3.

    Restriction on ownership: ownership must be dispersed among at least 100 shareholders, the five biggest of which may not own more than 50% of total shares outstanding. This provision calls for diffused ownership, making it difficult for large blockholders to acquire ownership stakes, and shareholders to form alliances to pose a takeover threat. Agency problems escalate under dispersed ownership (Ghosh and Sirmans 2003).

Do the monitoring benefits of high dividend payout offset the potential agency costs imposed by the restrictions on investment options and ownership structure? Ghosh and Sirmans (2003) find that outside directors enhance REIT performance, albeit weakly, but institutional owners have no impact. CEO ownership has a significantly negative impact on performance. The authors conclude that CEO has greater influence than outside directors on board composition and performance. Ghosh and Sirmans (2005) find that non-affiliated board members positively impact CEO compensation. This suggests that outside board members are aligned with the CEO, an indication that entrenched CEOs nominate directors that are loyal to him. Overall, the evidence suggests that weak monitoring allows the REIT CEO to be entrenched.

Ghosh and Sirmans (2006) examine the relation between managerial entrenchment and dividend policy of REITs. Similar to previous authors, they confirm the positive impact of entrenchment on dividend payout and attribute the result to the substitution hypothesis. They note that despite the high payout, large depreciation deductions leave considerable free cash flow at REIT managers’ disposal.Footnote 9 Hence, following Pan (2006), shareholders have the incentive to surrender power to the CEO through weak governance to induce him to return excess cash through dividend payments. As noted earlier, the market for corporate control is non-existent among REITs. As such, REIT managers are more interested in minimizing internal monitoring. CEO can garner influence over internal governance through length of service, role in the board, and participation in nomination committee.

Research Design and Hypothesis

In this paper, we analyze how CEO influence on internal governance mechanisms impacts dividend policy. To that end, we develop better proxies for CEO entrenchment. The standard proxies of CEO entrenchment include CEO tenure and CEO duality as the chairman of the board. Long tenure implies that based on past performance, the board has decided to retain the CEO. Reappointment accrues more power to the CEO, forcing outside directors to remain loyal to him for continuity of their service. Gradually, the board loses independence and the cost of replacing the CEO increases. Similarly, when the CEO holds chairmanship of the board, he sets the agenda and can influence directors’ voting decisions. Hence, companies with long CEO tenure and CEO duality are subject to high agency costs. Previous authors ignore the potential interaction effects between these variables. We create a composite index to capture various levels of entrenchment from the most to the least entrenched CEO.

To measure CEO influence on outside board members, we focus on CEO involvement in the nomination committee. In the absence of nomination committee, CEO has complete control over the slate of nominees. The next level of involvement is where the CEO is a member of the nomination committee, which allows him considerable power over the selection process, but also gives outside directors some negotiating power. The most independent nomination committee is the one that does not include the CEO. This structure enables outside directors to bring pressure on the CEO to accept their nominees in exchange for personal benefits (Hermalin and Weisbach 1998). Several recent studies explore the impact of CEO involvement in director selection on board independence. Shivdasani and Yermack (1999) report that market reaction to appointment of independent directors is significantly lower when the CEO is a member of the nomination committee. The authors conclude that entrenched CEOs use their involvement to reduce pressure from active monitoring. Gerety et al. (2001) find that investors react negatively to adoption of long-term incentive plans for directors when the CEO is involved in director selection. Vafeas (1999) finds that the composition of nominating committee has no influence on the number of outside directors, but it affects their independence. Ruigrok et al. (2006) find that firms with nomination committees have a higher number of independent directors. Feng et al. (2005) demonstrate that CEO’s influence on the board is partly determined by his role in the nomination committee. Feng et al. (2007) conclude that when the CEO is not involved in the nomination process, directors receive higher equity-based compensation as an incentive to monitor.

Based on the theoretical and empirical evidence and the regulatory structure of REITs, we develop the following hypothesis:

The most (least) entrenched CEOs have the longest (shortest) tenure and hold (do not hold) dual positions of CEO and board chair. REITs with the most (least) entrenched CEOs and with no nominating committees (with nominating committee where the CEO is not a member) pay the highest (lowest) dividends.

Both substitution and optimal entrenchment hypotheses predict a positive relation between managerial entrenchment and dividend payout. Under the substitution hypothesis, dividends and monitoring mechanisms are substitutes. Entrenched managers pay dividends to avoid cash stockpiles which can provoke shareholder sanctions ultimately leading to removal of incumbent management (John and Knyazeva 2006). The optimal entrenchment hypothesis asserts that shareholders willingly surrender power to CEOs to protect against unfriendly bidders. Protected CEOs are more likely to pay dividend instead of using the cash on value-destroying acquisitions. The contradictory findings for USA and international markets suggest that managerial behavior is influenced by the protection and enforcement of shareholder rights in the market, not by firm level agency conflicts. High dividend payment by REITs reassures shareholders that managers’ opportunity to expropriate wealth is limited. Weak market for corporate control implies that REIT managers are less concerned about the takeover risks of cash accumulation. Managers’ investment opportunities are also limited by law. In this environment, support for our hypothesis implies: (1) although hostile takeover attempts are rare, entrenched REIT CEOs are still reluctant to stockpile cash; (2) despite restricted investment choices, shareholders seek to maximize dividends by allowing managers extra protection through weak internal governance.

Data and Summary Statistics

We identify all exchange listed REITs from the SNL database and the NAREIT annual lists for the years 1999 and 2000. The number of exchange-traded REITs for these years is about two hundred. The data on financial variables are collected from the SNL database. The data on CEO compensation, and board characteristics are collected from proxy statements. The non-availability of proxy statements limits our dataset to a total of 236 REITs, 118 for each year.

The summary statistics are presented in Table 1. For 1999 and 2000, the average payout of REITs over the previous 5 years is 144% of net income. The high payout ratio is attributable to the mandatory payout requirement of REITs to maintain tax-exempt status. Although payout ratio is high, REITs still have considerable discretion over payout as a proportion of funds from operations (FFO). For instance, REITs pay out, on average, only 68% of FFO over the previous 5-year period.

Table 1 Descriptive statistics for dividend payout, CEO entrenchment, CEO ownership, board monitoring and firm characteristics of REITs: 1999 and 2000
Full size table

Following previous studies, we employ CEO tenure and CEO duality as proxies for CEO entrenchment. On average, CEOs have served 6 years in their current positions, and about half (53%) of the CEOs serve in the dual role as the chairman of the board of directors. Equity-based compensation and stock ownership are widely used to align CEO’s personal incentives with shareholders’ interests. CEOs in our sample receive about 43% of compensation in the form of stock options, phantom stocks and/or restricted stocks. Based on a sample of REITs from 1997–2000, Pennathur et al. (2005) report that average stock-based compensation accounts for a little over 50% of total compensation. Board composition can also help to mitigate agency costs. Previous studies show that smaller boards (Yermack 1996) and a higher proportion of independent directors are more effective in monitoring. Non-independent directors include individuals that are current or retired managers and their relatives, and outside directors who receive payments from the company in excess of the compensation for directorship. The other outside directors are labeled independent. On average, REITs have eight board members, 59% of whom are independent.

In Table 1, we report the summary statistics on several control variables including firm size, performance, growth, and leverage. Return on assets, measured as the ratio of earnings before interest and taxes to total assets, average 4% over a 5-year period. REITs enjoy an average of 22% annual growth in earnings and 30% annual growth in total assets over the 5-year period, have market-to-book ratio of 1.16, and debt to total assets ratio of 46%. Real estate assets account for over 94% of total assets, possibly an effect of the restriction on investment options open to REITs. As an asset class, cash represents only about 2% of total assets, a result of regulatory provision on payout. Seventy-three of the 236 REITs have nomination committees, 40 of which are independent—CEO is not a member of the committee.

Univariate Analysis

CEO Entrenchment and Dividend Policy

To investigate the impact of agency costs on REIT payout policy, we sort the sample according to the level of CEO entrenchment. The longer the CEO serves in the position, the more influence he has on the board, and the more entrenched he is. If CEO serves as the chairman of the board, he has the power to set the agenda and pressure the other members to vote according to his preference. We classify the REITs where the CEO has been appointed for more than 4 years (median for the sample) and also serves as the chairmen of the board as the high CEO entrenchment group. REITs where the CEO has a short tenure (less than 4 years) and does not serve as chairman of the board comprise the low CEO entrenchment group. The rest of the firms are classified as the average CEO entrenchment group. Fifty-nine of the 236 observations have low CEO entrenchment, 102 have average CEO entrenchment and 75 have high CEO entrenchment.

In Table 2, we report the statistics for dividend payout, alternative proxies of agency cost, and financial and performance variables for the three CEO entrenchment levels. Consistent with our basic hypothesis, the high entrenchment group pays out 60% more dividends as a percentage of earnings than the low entrenchment group. Several other attributes of the high and low entrenchment groups deserve note. The average tenure of high CEO entrenchment group is 7.5 years longer than the low CEO entrenchment group. The high entrenchment firms invest more in real estate assets which is attributable to the regulatory restriction in investment choices. Clearly, the absence of hostile acquirers from outside the real estate sector help entrenched CEOs remain insulated. The firms with entrenched CEOs have less cash available relative to total assets, and lower growth rate of assets. Harford et al. (2006) also find that firms with entrenched CEOs hold less cash because they dissipate the cash on value-destroying investments to avoid cash stockpile.

Table 2 Financial and operating characteristics of REITs by CEO entrenchment
Full size table

Composition of Nomination Committee and Dividend Policy

Firms with nomination committees should have more independent boards. To test this proposition, we conduct univariate analysis separately for REITs with (Table 3) and without (Table 4) nomination committees for the three levels of entrenchment. The results are reported in Tables 3 and 4. CEO entrenchment has no impact on dividend policy among REITs with nomination committees. For REITs with no nominations committees, CEO tenure and CEO duality are significantly higher for the high CEO entrenchment group. Conceivably, to minimize monitoring and protect his tenure, CEO controls the nomination process to elect directors loyal to him. These REITs pay out 185% of earnings when CEO is strongly entrenched compared to a significantly less 109% when CEO is less entrenched. Assuming boards without nominating committees are less independent, these patterns support the contention that entrenched CEOs pay high dividends to appease shareholders. Furthermore, when CEO entrenchment is low, firms with nomination committees pay out more dividends (147% of the earnings) compared to firms without nomination committees (109% of the earnings). If firms with nominating committees have more independent boards, this result demonstrates that when CEO is less entrenched, directors force managers to disgorge cash to shareholders.

Table 3 Financial and operating characteristics of REITs with and without nomination committees (with nomination committee)
Full size table
Table 4 Financial and operating characteristics of REITs with and without nomination committees (without nomination committee)
Full size table

In Table 5, we report the 5-year average payout for low, middle and high CEO entrenchment levels. For each entrenchment level, we report average payout for high and low levels of the other variables, and p values for difference of means between the subgroups. The difference in mean dividend payout between high and low entrenchment levels is significant for most of the other variables. In contrast, difference in payout between the high and low levels of the other variables is significant in very few cases. To elaborate, among governance and CEO incentive variables, firms with high CEO entrenchment pay out more dividends for both levels of CEO ownership, board size, and board independence. Dividend payout is also significantly higher in high CEO entrenchment group for both levels of growth in total assets, growth in earnings, and investment in real estate. Furthermore, CEO entrenchment has stronger impact on dividend policy for firms with poor performance (in terms of 5-year average ROA and market-to-book ratio). Contrary to our expectation, there is no significant difference in the dividend payout of low and high growth firms for any level of entrenchment. Firms with high leverage pay out more dividends, suggesting that monitoring through cash distribution to bondholders and stockholders are complementary to each other. Overall, these results support our contention that firms with high agency costs, indicated by high CEO entrenchment, pay out more dividends. The positive relation between CEO entrenchment and dividend payout level is consistent with both substitution and optimal entrenchment hypotheses.

Table 5 Dividend payout (5-year average) classified by CEO entrenchment and other characteristics
Full size table

Multivariable Regressions

We interpret the univariate results to be consistent with the notion that entrenched managers pay high dividends to deploy free cash flow, and avoid cash stockpiling which may incite investors. In this section, we focus on the impact of CEO entrenchment—CEO tenure and duality, and participation in the nomination process—on internal governance and dividend payout. Long tenure, dual role as chairman of the board, and involvement in director selection combine to result in a captive board with outside directors keen to oblige the CEO to win reappointment. The powerful CEO can pressure a weak, condescending board to acquiesce to his demands of excessive perquisites and other personal benefits. In exchange, the CEO uses his influence on the nominating committee to get his preferred (loyal to him) candidates appointed to the board. The entrenched CEO pays high dividend to win shareholders’ confidence and trust in the management and the board. Shareholders that are content with dividends are unlikely to align with dissidents who may attempt to overthrow the management.

We model dividend payout in a multiple regression framework to capture the potential interaction between the proxies for monitoring and agency cost, CEO involvement in director selection, and the standard control variables. To incorporate the effect of nomination committee, we estimate the model separately for REITs with and without nomination committees. The “with nomination committee” group includes the cases where the CEO is a member of the committee. Following Gerety et al. (2001), we argue that when the firm has a nomination committee, outside directors have the opportunity to negotiate with the CEO over the composition of the board, even if the CEO is a member of the committee. They are denied that opportunity when there is no nomination committee. Accordingly, CEO is most entrenched when the firm has no nomination committee because the CEO has complete control over director selection.

We develop a CEO entrenchment index to represent high, low and average levels of entrenchment, using CEO tenure and CEO duality as proxies. The impact of these variables on entrenchment is uniformly positive, but within an individual firm, they can reinforce or offset each other depending on respective levels. A composite index can capture interaction effects. Entreind1 is a dummy variable with value 1 for firms with CEO tenure of more than 4 years and the CEO in a dual role as the chairman of the board, and 0 otherwise. This is the most entrenched group of CEOs, with the highest agency cost. Entreind2 is a dummy variable with a value 1 for firms whose CEOs are in the position for less than 4 years and serve as chairmen of the board, and for firms whose CEOs have more than 4 years of service but do not serve as Chairmen of the board, and 0 otherwise. This is the group with average entrenchment in that the adverse effect of one factor is partly offset by the favorable effect of the other. Entreind3 is a dummy variable with a value 1 for firms with short CEO tenure (less than 4 years) and CEO not serving as the Chairman of the board. This group has the lowest agency cost. We do not include CEO ownership in the index because its effect on entrenchment is non-linear. At low ownership levels, managers identify with shareholders’ interest; but, beyond an optimum level of ownership, managers’ excessive power and influence results in entrenchment.

The maintained hypothesis is that CEO entrenchment and weak governance by the board lead to high dividend payout. The proxies of agency cost include the CEO entrenchment index (positive impact on dividend payout), CEO equity-based compensation (negative effect), CEO ownership (negative at low levels, positive at high levels), board size (positive effect), and proportion of outside independent directors (negative effect). While the predicted effect of CEO characteristics on dividend policy is straightforward, the effect of board composition is more complex. A firm with a powerful CEO and a weak board pay high dividends to appease shareholders so that the CEO and loyal board members can continue to consume the private benefits of power. Larger boards and boards with a small number of independent directors are weak monitors, hence a positive coefficient for board size, and negative coefficient for percentage of outside director are predicted. The firms with better performance and higher growth pay lower dividend to fund investment. Accordingly, 5-year average ROA, and growth in total assets should have negative coefficients. The ratio of debt to total assets, and cash to total assets are proxies for free cash flow at CEO’s disposal. Higher leverage reduces free cash flow. So, leverage is an alternative monitoring mechanism to dividends implying a negative coefficient for leverage ratio. High cash to total assets ratio indicates high free cash flow, hence cash should have a positive impact on dividends. Finally, larger firms are more stable with limited investment opportunities, hence they need less capital. This predicts a positive coefficient for firm size.

The results are reported in Table 6. Entreind3 which identifies the least entrenched group of CEOs, with the lowest agency cost, represents the control group. The direct relationship between CEO entrenchment and payout predicts positive coefficients for the two included entrenchment dummies. Three models are estimated, model 1 for the whole sample, and models 2 and 3 are for REITs without and with nomination committees, respectively. In all models, we include the two entrenchment indexes, Entreind1 and Entreind2, and independent variables representing monitoring, performance, growth, leverage and other control variables. The coefficient on ROA is significantly negative in all models which illustrates that better performing REITs pay lower dividends to conserve funds for investment. The negative impact of CEO equity compensation suggests that long term compensation aligns CEO incentive with shareholders, mitigating agency costs. Accordingly, dividends are lower. The dummy variables for CEO entrenchment levels are significantly positive which supports the hypothesis that entrenchment leads to higher dividends. The negative sign on board size is contrary to the notion that larger boards have coordination problems resulting in weak monitoring.

Table 6 Results from multivariable regressions for dividend payout
Full size table

In model 2 which includes REITs without nomination committees, the results are similar. CEO has strong influence on director selection in firms without nomination committees. Such involvement leads to entrenchment, and results from model 2 demonstrate that entrenched CEOs pay high dividends. This result is consistent with Hu and Kumar’s (2004) finding that CEO tenure and CEO duality have positive impact on dividend payout ratio. CEOs with long tenure are likely to be more entrenched. As the chairman of the board, CEO can influence the nomination and election of directors, such that outside directors remain loyal to him, making him powerful and entrenched. Entrenchment of managers increases agency costs. Keen to protect their positions, entrenched CEOs pay higher dividends to avoid shareholders’ sanctions (Zwiebel 1996). This result is particularly intriguing because mandatory dividend payment for REITs is high by regulation. However, because of special regulation on investment options and ownership structure, alternative governance mechanisms are not as effective for REITs. Our results show that even if external governance is weak, and dividend payout is high by regulation, entrenched REIT managers pay higher dividends. Following Harford et al. (2006) and John and Knyazeva (2006), the environment of strong protection and enforcement of shareholder rights, and performance oriented incentive structure induce managers to return excess cash to shareholders rather than accumulating it.Footnote 10 As Harford et al. (2006) and John and Knyazeva (2006) observe, even the most entrenched managers in the USA are wary that the visibility of excess cash and the associated agency costs may cause shareholder dissent, and ultimately removal from power.

Model 3 includes REITs with nomination committees, where directors are more independent. In contrast to models 1 and 2, the effect of CEO entrenchment is not significant in model 3. Clearly, the results for the whole sample are driven by REITs without nomination committees. Model 3 demonstrates that CEO entrenchment does not constitute a barrier to monitoring if the CEO has limited involvement in director selection. Further, once we incorporate the effect of the composition of the nomination committee, other proxies for CEO entrenchment (i.e. tenure and duality) have no impact. This paper is the first to document the important role of the composition of the nomination committee on CEO entrenchment and internal governance. Finally, the significant and negative coefficient of outside directors is consistent with the notion that firms with stronger boards pay out lower dividends.

To further explore the role of nomination committee in internal governance, we refine the CEO entrenchment index in terms of the extent of CEO involvement in director selection. Specifically, we create six dummy variables, three for each of the two classes of REITs, with and without nomination committees. The three groups within each class represent the three levels of CEO entrenchment defined earlier. The group with the most highly entrenched CEO (tenure more than 4 years and in dual role as chairman, Entreind1 in the previous specification) and without nomination committee is chosen as the control group. Under optimal entrenchment hypothesis, this group pays the highest dividend, which predicts negative signs for the other five dummy variables. We estimate four regression models. In model 1, we focus on CEO entrenchment levels with the five dummy variables. Model 2 highlights other CEO characteristics, model 3 includes board characteristics, and model 4 includes all the variables.

The results are reported in Table 7. Consistent with our hypothesis, entrenchment dummies are all significantly negative in models 1 and 4 which confirm that REITs with the most entrenched CEOs (long tenure, duality, and no nomination committee) pay the highest dividend. The coefficients for 5-year average ROA and growth in total assets are significantly negative, highlighting that firms with superior performance and more growth options pay lower dividends. In model 2, we add CEO equity-based compensation and CEO ownership. The coefficient for CEO ownership is negative and significant, which corroborates the notion that CEO ownership helps align CEO’s interests with those of shareholders, leading to lower agency costs, and lower dividends. The absence of a significant relationship between CEO equity-based compensation and dividend policy suggests that compensation mix does not substitute for dividends as monitors. Model 3 controls for the impact of board composition. While large boards are associated with lower dividends, outside directors have little impact on dividend policy. The weak impact of board composition on firm performance and dividend policy is consistent with the implications of Hermalin and Weisbach’s (1998) model that board structure evolves as a compromise between independent directors and the CEO, such that in equilibrium, board attributes are irrelevant. The results in model 4 reveal similar patterns.

Table 7 Results from multivariable regressions for dividend payout
Full size table

In summary, to protect and enhance their careers, entrenched CEOs pay higher dividends to appease shareholders. CEO stock ownership helps align CEO’s personal goals with shareholders’ interests, reducing the need for discipline through high payout. Finally, CEO equity-based compensation and board composition have little impact on dividend policy. The evidence is also consistent with residual theory that firms with superior performance and growth opportunities pay out less cash. The most significant finding is that CEO entrenchment is mainly determined by CEO’s involvement in director selection through the nomination committee. Once the composition of the nomination committee is included as an explanatory variable, the standard proxies for entrenchment lose significance. To our knowledge, this study is one of a small group to explore the important role of nomination committee in internal governance.

Discussion and Interpretation

The finding that entrenched CEOs pay higher dividends is consistent with the result in other studies using US data. However, there are important differences between non-regulated firms and REITs. One, it is argued that agency conflicts are mitigated by regulation. It is not apparent to what extent this generalization applies to REITs because the purpose of regulation of REITs is to enhance transparency and diversification benefits, not to set fair rates of return or enforce fiduciary responsibility. Two, by regulation, REITs must pay out nearly all taxable earnings as dividends. However, because of large depreciation deductions, REIT managers still have large free cash flow at their disposal. Three, REIT regulation may weaken the external governance of the takeover threats. In conjunction, these unique aspects raise the interesting question of whether shareholders offer managers any inducements in the form of weak governance to extract more cash than is mandated by regulation. Our analysis reveals that the answer to this question is: yes, managers with weak internal monitoring pay significantly higher dividends. Clearly, shareholders are wary of managers’ access to free cash flow, notwithstanding the fact that REIT managers have much less discretionary cash flow than non-regulated firms.

Both substitution and optimal entrenchment hypotheses assert that self-interested managers wish to continue enjoying the special privileges associated with power and prestige, but feel threatened by the active takeover market in the USA. Entrenched managers’ prefer to accumulate cash for private benefits and consumption. Under the substitution hypothesis, entrenched managers pay dividends to dissipate cash to escape shareholder outrage and pressure from institutional owners to reduce the agency costs of cash stockpiles. However, it is not clear why managers would use cash for dividend distributions over takeover-resistant actions (i.e. share repurchase, acquisitions). In the optimal entrenchment model, shareholders surrender power through weak governance to allow managers entrenchment and protection against hostile takeovers. Protected managers pay higher dividends. In REITs, hostile takeovers are rare and investment opportunities are limited by regulatory restriction, allowing entrenched managers greater opportunity to usurp corporate assets for private benefits. That REITs on average pay higher dividends than required by law is consistent with the argument that strong protection and enforcement of investors’ rights force US managers to abide by shareholders’ preference for dividends when the firm has limited investment opportunities. Even in an environment such as REITs where takeover market is perceived to be weak, managers seem keen to avoid shareholder sanctions and ultimate removal. Harford et al. (2006) suggest that the strength of country-level shareholder rights dominates the effect of firm-level variation of control of agency conflicts. Our evidence indicates that the dominance transcends entire industry sectors.

An interesting finding from our analysis is that the strength of internal governance, especially CEO influence on director selection and board structure, significantly impacts dividend policy. Our analysis clearly demonstrates that the more entrenched is the CEO due to weak internal governance, the higher is the dividend payout. If the amount of dividend payout is determined by the vulnerability of the CEO to the market for corporate control, our evidence indicates that the CEO feels more protected under weak internal governance. In essence, even in a weak market for control, CEO capitalizes on opportunities for entrenchment. If these opportunities require shareholder vote, CEO pays higher dividends to influence shareholders. This is consistent with Pan’s (2006) model.

Finally, do the monitoring benefits of dividend payment justify the costs of managerial entrenchment? The costs of managerial entrenchment are well documented in the literature. For evidence, our analysis reveals that performance and growth of REITs that pay higher dividends is significantly inferior. Ghosh and Sirmans (2003) report that performance deteriorates under suboptimal governance. Ghosh and Sirmans (2005) find that CEO receives higher compensation when monitoring is weak. Harford et al. (2006) show that entrenched managers dissipate excess cash on value-destroying investments to deter hostile takeovers. Anecdotal evidence suggests that an entrenched CEO, in collusion with a condescending board, extracts rent for personal benefit. However, managerial entrenchment has potential benefits too. For example, job security may encourage managers to evaluate projects from a long-term perspective, instead of adopting a short-term, myopic outlook. Allowing CEO entrenchment to induce high dividends is rational so long as the net costs of entrenchment do not exceed the adverse value consequences of investment in takeover-resistant initiatives, and the cost of ultimately removing errant managers.

Conclusion

Staring with Black (1976), the dividend puzzle has intrigued financial economists for long. One idea that has received consistent support from theoretical and empirical literature is that dividends can be effective in monitoring managers, and mitigating agency costs. Dividends alleviate agency conflicts by reducing managers’ access to free cash flow. In addition, dividend paying firms have to raise investment funds more frequently from the market, subjecting managers to the discipline of investment bankers and institutional investors. It follows that entrenched managers who want to minimize monitoring will pay less dividends such that managerial entrenchment and dividend payment should be negatively related.

The empirical evidence based on United States data is contrary to this proposition. The international evidence is consistent with it, however. Researchers have attributed these conflicting results to the strong corporate governance in the USA. Specifically, the strong protection and enforcement of shareholder rights in the USA force entrenched managers not to expropriate corporate resources or stockpile cash (by not paying dividends). Such tendencies can provoke shareholder sanctions, and ultimately to transfer of power. Our objective is to examine this notion with REITs, particularly from the perspective of the special regulatory provisions.

The REIT environment which can insulate managers from governance by external agents (institutions, investment bankers, and market for corporate control) is the ideal setting to examine how CEO entrenchment shapes dividend policy. We focus on the internal governance mechanisms. We set up models of dividend payout as a function of proxies for agency costs of managerial entrenchment including existence of the nomination committee, CEO tenure, CEO duality, equity-based compensation, stock ownership, board size and percentage of outside directors. We develop an index of entrenchment incorporating CEO tenure, CEO duality, and composition of nominating committee. Based on the fact that REIT managers have access to considerable amount of free cash flow despite the high dividend payout, our central hypothesis is that entrenched managers pay higher dividends to insulate them from the disciplining sanctions of skeptical shareholders, and discourage hostile suitors.

The analyses reveal two noteworthy findings. One, segregating the overall sample in terms of CEO entrenchment levels reveals that highly entrenched CEOs pay significantly higher dividends. Second, further analysis in terms of various levels of alternative monitoring proxies confirms that CEO entrenchment and existence of nomination committees are the only two significant factors in explaining differential dividend payout ratios. Most importantly, CEO entrenchment dominates the effect of other governance variables. This evidence is consistent with the substitution hypothesis that dividends and other governance mechanisms act as substitutes. The result also suggests that even if the external governance system for the sector is less effective, the pressure from the overall market for corporate control in the USA forces REIT managers to payout dividends to shareholders rather than expropriate assets. This corroborates the view of previous authors that overall governance in a market, not the firm or industry level agency conflicts, is what influences managerial behavior. Finally, our finding that shareholders allow managers to influence board structure through nomination committee bears on the optimal entrenchment model.

The costs of managerial entrenchment are well documented in literature. These include wealth expropriation by entrenched managers, and investment in takeover-resistant, value-destroying projects. We find that CEO entrenchment adversely impacts performance and growth, and enhances CEO compensation. Why do shareholders allow entrenchment despite these costs? We have identified several sources of value gains. If entrenched managers feel secure in their jobs, they may focus on long-term initiatives, rather than short-term, myopic decisions. Protected managers are expected to pay dividends before investing in non-value-maximizing projects. In essence, some entrenchment may be necessary to induce managers to act to protect and enhance shareholder interest. This will save the cost of monitoring and ultimately removing them. These potential gains may more than offset the costs of entrenchment.

It is noteworthy that REITs are moving towards establishing nomination committees and reducing CEO involvement in director selection during the study period. In 1999, only 31 out of 118 REITs had nomination committees, and 16 of these committees were composed of independent board members. In 2000, 42 REITs had nomination committees, 24 of which were independent. Under the optimal entrenchment model, reducing managerial entrenchment may not always be value-maximizing. Future research can explore if alternative mechanisms develop to offer protection to self-interested managers.