There are instances where CEO turnover occurs, even if the company has not made any significant strategy changes, and the new CEO possesses similar abilities as the predecessor. This paper aims to provide a rational explanation for this seemingly irrational phenomenon. One possible reason for this “aggressive” CEO turnover is the board’s desire to reduce the information rents earned by the privately informed CEO. Specifically, the incumbent CEO has a temptation to “sandbag” the board about profitability prospects to secure more generous incentive pay for future implementation, and a (seemingly aggressive) replacement policy helps discourage this kind of gaming. That is, instead of “information-based entrenchment” as suggested by the literature (Laux 2008; Inderst and Mueller 2010), this paper shows a countervailing effect that the CEO’s private information (combined with the later-stage moral hazard problem) may lead to her dismissal more often than the ex post efficient benchmark.

We introduce real decisions (a project choice decision, an investment scale decision, and an information acquisition decision) to the Dye (1985) voluntary disclosure framework and examine how the prospect of voluntary disclosure affects managers' real decisions. Riskier projects lead to more volatile environment and hence entail higher efficiency loss at the subsequent investment scale decision stage if managers are uninformed. If managers are informed, they can withhold bad information, and the value of this option is higher for riskier projects. We show that the voluntary nature of managers' disclosure may lead to two types of inefficiencies: (1) managers may choose riskier projects, which generate lower expected cash flow due to the higher efficiency loss at the subsequent decision stage, and (2) managers may over-invest in information acquisition, because informed managers with bad information have the option to pool with uninformed managers and benefit from being overpriced.

We study a corporate board tasked with monitoring a firm's CEO and providing incrementally decision-relevant information. The board has both compensation and non-pecuniary incentives—we label the latter board bias. Friendly boards have muted information gathering incentives, but can more effectively engage in cheap talk communication with management. As a result, the direction of the optimal board bias is determined by the CEO's initial information advantage: the board should be weakly friendly if the CEO is endowed with precise information, and weakly antagonistic (to the CEO) otherwise. Aside from assembling a friendly board, another way for shareholders to foster CEO/board communication is by granting the CEO more equity. In general, we find board friendliness and CEO equity grants to be positively associated, in equilibrium. This provides an optimal contracting rationale for an empirical regularity often interpreted as friendly boards facilitating rent extraction.