Employment capital, board control, and the problem of misleading disclosures *.

The extensive literature on CEO power suggests a concern for maintenance of a "healthy" balance of power between the CEO and the board (e.g., Daily and Schwenk, 1996; Morse, 2002). This focus is not a theoretical abstraction. Empirical evidence suggests that high CEO power is associated with lower firm performance (Pearce and Zahra, 1991), dominance of the strategic agenda (Golden and Zajac, 2001), and investment distortions (Mahoney et al., 1997). The primary reason for these findings seems to be that, in general, high CEO power leads to executive entrenchment (Sundaramurthy et al., 1997) and the possibility of what Diamond (1993) referred to as "control rent appropriation," or the accumulation of excessive perks or other benefits of office.

Indicia of power are varied in the literature, but among those commonly utilized are tenure (Finkelstein and Hambrick, 1989; Ocasio, 1994; Shen and Cannella, 2002) and managerial ownership (Finkelstein and Hambrick, 1989; Pollock et al., 2002; Sundaramurthy et al., 1997). Tenure is seen as increasing the power of the CEO via intra-organizational influence and authority. In turn, oversight by the board or other monitors may be less effective and the CEO's job security is likely to increase, at least where threshold expectations are met (Ocasio, 1994; Shen and Cannella, 2002). Consistent with this line of analysis, Williams et al. (2005) report a positive relationship between managerial tenure and illegal activity. Therefore, longer-serving, and thus more powerful, CEOs are more likely to be able to leverage their standing and relative independence of action by issuing misleading disclosures designed to give the appearance of better than actual performance. Doing so permits these CEOs to maintain their power base and freedom from oversight by continuing to deliver results (even though ephemeral).

Hypothesis 2: The incidence of misleading disclosures will increase as CEO tenure increases.

Agency theory suggests that equity ownership facilitates incentive alignment and therefore will suppress purely self-interested actions by managers (Jensen and Meckling, 1976; Pollock et al., 2002). This is undoubtedly true where managers own little or no stock and begin to accumulate a position of some substance. But as executive ownership increases from minimal levels, the question remains whether, at higher levels of ownership, incentive alignment benefits from incremental increases in ownership outweigh the possibility of entrenchment. In other words, for CEOs who already own a significant block of stock, will adding more increase their incentives or only their voting power? Empirical evidence, in fact, suggests a curvilinear relationship in which any incentive alignment benefits accrue at low levels of ownership but level off and decline as more equity is accumulated (McConnell and Servaes, 1990; Morck et al., 1988). Thus, in terms of the implications of CEO power, higher levels of ownership may increase the likelihood of entrenchment (Finkelstein and Hambrick, 1989) and lead to lower firm returns (McConnell and Servaes, 1990; McWilliams, 1990).

Applying this logic to the case of misleading disclosures leads to the conclusion that either low or high levels of CEO ownership will be associated with a greater likelihood of misleading disclosures. At low levels of ownership, executive interests are not aligned with those of shareholders (Jensen and Meckling, 1976) and the executive lacks any power base associated with ownership that might permit him or her to avoid questions concerning less than expected performance. By contrast, at higher levels of ownership, power provides the means to mislead and greater equity holdings provide the incentive to boost results and (presumably) share prices (Finkelstein and Hambrick, 1989; McConnell and Servaes, 1990).

Hypothesis 3: There is a curvilinear relationship between CEO stock ownership and the production of misleading disclosures, such that misleading disclosures are more likely to occur at either lower or higher levels of CEO stock ownership.

The Moderating Role of the Board

Recent research suggests that boards are exercising authority more directly and becoming more involved in firm activities and decision making (Buchholtz et al., 2005; Golden and Zajac, 2001; Westphal and Fredrickson, 2001; Zahra et al., 2005). This development is consistent with the board's legal duties and with agency theory, both of which assert that the board's primary responsibility is that of oversight and control (Deutsch, 2005; Fama, 1980; Walsh and Seward, 1990). Key to the board's ability to effectively oversee and control managerial activities, however, is its independence from management (Daily and Schwenk, 1996; Johnson et al., 1996). According to Johnson et al., "Directors who are personally influenced by the CEO ... may be less effective monitors of firm management ... [and therefore] boards comprised predominately ... of independent directors are expected to more effectively monitor management self-interest" (1996: 416). Thus, the proportion of unaffiliated outsiders on the board typically has been associated with improved monitoring (Daily and Schwenk, 1996; Johnson et al., 1996; see also Johnson et al., 1993). Consistent with agency theory, equity ownership by directors also is posited to increase the likelihood that boards will actively monitor the behavior of management (e.g., Hoskisson et al., 1994; Johnson et al., 1996; Johnson et al., 1993).

Average director tenure also may serve to increase the board's independence and authority, in part by reducing the immediacy of any sense of obligation to the CEO (Boeker, 1992; Wade et al., 1990). Director tenure facilitates the development of a richer base of organizational knowledge (Fiske and Taylor, 1991; Golden and Zajac, 2001), permitting the board to reach its own conclusions rather than relying upon management for explanation. Finally, as the board serves for a longer period of time, it has the opportunity to develop, and to be seen, as a separate source of power and authority within the firm. Thus, increasing tenure is likely to empower the board and to reduce the likelihood that misleading disclosures will be issued.

The relationship between the board and management implies that the board's ability to monitor and control CEO behavior will decrease the likelihood that management will succeed in issuing misleading disclosures. Thus, outside directors, director stock ownership and director tenure, all indicators of board independence, will act as moderating variables. Where an independent board can exercise sufficient direct control, managers are less likely to attempt to mislead, but where board control is weak, we can expect that managers will be increasingly likely to issue misleading information.

Hypothesis 4a: As board control increases, the relationship between firm performance and the incidence of misleading disclosures will decrease.

Hypothesis 4b: As board control increases, the relationship between CEO power and the incidence of misleading disclosures will decrease.

METHODS

Sample and Data Collection

We constructed the sample from an electronic word search in Lexis/ Nexis Business News that identified reports of financial restatements involving legal action or SEC inquiries into overly-aggressive or misleading accounting practices, particularly those in which revenues or earnings were overstated or inventory or debt was understated. Cases involving purely technical corrections or those in which upward revisions in sales or earnings were undertaken in accordance with generally accepted accounting principles (in other words, cases in which the original reports were less positive than justified) were excluded from the sample. We selected the years between 1996 and 1999, inclusive, as the timeframe for the study. These years corresponded with generally increasing economic activity and market performance, during which expectations were high and steadily rising. The years from 2000 to the present were excluded because of the advent of the bear market beginning in approximately March 2000, the effect of which may have been to change the standards of performance evaluation applied to firms and their managers, and therefore to alter managerial incentives.

We then sought to identify matching organizations based upon each primary sample firm's four-digit SIC code, total assets, and number of employees. This design controls for inter-firm and extra-organizational influences, specifically those relating to shared industry and environmental effects and size-related advantages or disadvantages. Data were matched based upon each firm's assets and employees in the year immediately preceding the primary sample's restatement period. Assets and employees were used as match criteria because these totals were less likely than earnings to have been subject to manipulation or misstatement. We then conducted a word search for the matching firms in order to ensure that they had not restated earnings. Matches were identified for all but two of the primary-sample firms, yielding a total sample of 140 companies, comprised of 70 restating firms and 70 non-restating firms. The same one-year lag procedure for data collection was also utilized in the matching process. Full data were gathered for each company (both restating and non-restating) for the year immediately preceding the restatement period.

Variables

As indicated, the dependent measure was the incidence of a restatement to prior years' earnings. This variable was coded "1" for firms who restated earnings and "0" for those who did not.