In a setting with moral hazard and optimal compensation contracts, dividends must be coordinated with leverage to maintain efficient investment incentives. Dividends are more flexible than debt, allowing firms to reduce the cost of financial constraints. Dividends, however, also provide managerial flexibility. With low payouts, managers enjoy excessive flexibility over investment choices, and pursue projects that increase their information advantage. This increases the rents they receive under the dynamically consistent compensation scheme. With more disciplined dividends, these furtive investments create a debt overhang problem, such that shareholder wealth is maximized with a compensation scheme that neglects low outcomes (at bondholder expense) and reduces the manager's rents. Investment and compensation choices, therefore, reflect both the flexibility of the financial policy and the flexibility in the manager's dividend choice. When dividends cannot effectively restrict managerial opportunism, short-term debt payments can substitute, but at the cost of default or exacerbating the shareholder-bondholder conflict. The optimal financial policy balances the benefits of financial flexibility and the cost of managerial flexibility. Our theory is closely related to the theory of DeAngelo and DeAngelo (2007), where leverage and dividends combine to provide optimal financial flexibility, while controlling for the agency costs of cash accumulation. Dividends, rather than debt payments, constrain managerial discretion as this preserves the ability to borrow and enhances the ability to issue equity (by enhancing the firm's reputation for returning cash to equity holders). In our analysis, the agency costs of cash accumulation are derived in an explicit and unique manner that operates through managerial information advantages. Disbursements reduce these costs, but may also exacerbate the shareholder-bondholder conflict, which is similar to Myers'(1977) debt overhang problem, but again unique in that it operates through managerial compensation. The optimal financial policy trades off the unique aspects of the agency and debt overhang problems, in contrast to DeAngelo and DeAngelo, where it trades off agency costs illustrated by Jensen (1986) with the security valuation costs illustrated by Myers and Majluf (1984)). Our theory also recognizes that since dividends are designed to control managerial discretion, it may be difficult to induce controlling managers to make disbursements. Short-term debt imposes a stronger commitment, but at the risk of additional costs, so that our financial policy also incorporates Jensen's concern with the relative softness of dividends, which is absent in DeAngelo and DeAngelo.