We study a setting wherein a divisional manager undertakes personally costly effort to improve the profitability of an investment project. The manager's choice of innovation effort is subject to a holdup problem because of the ex post opportunism on the part of headquarters. We analyze and contrast the performance of centralized and delegated forms of investment decision-making. We find that delegation improves the manager's innovation incentives. We identify conditions for each of the two organizational forms to emerge as the optimal choice, and relate these conditions to characteristics of firms' investment opportunity sets.