This thesis consists of two essays that examine several problems in corporate finance and mechanism design. The central theme is endogenous agency conflicts and their impact on dynamic investment decisions. The first essay features auctions of assets and projects with embedded real options, and subsequent exercises of these investment options. The essay shows timing and security choice of auctions endogenously misalign incentives among agents and derives the optimal auction design and exercise strategy. The second essay studies implications of endogenous learning on irreversible investment decisions, in particular, how learning gives rise to asymmetric information between managers and shareholders in decentralized firms. Depending on the quality of the project, the optimal contract between principal and agent distorts investments in ways that has not been examined in the literature. Specifically, in Chapter 1 of the dissertation, I study how governments and corporations auction real investment options using both cash and contingent bids. Examples include sales of natural resource leases, real estate, patents and licenses, and start-up firms with growth options. I incorporate both endogenous auction initiation and post-auction option exercise into the traditional auctions framework, and show that common security bids create moral hazard because the winning bidder's real option differs from the seller's. Consequently, investment could be either accelerated or delayed depending on the security design. Strategic auction timing affects auction initiation, security ranking, equilibrium bidding, and investment; it should be considered jointly with security design and the seller's commitment level. Optimal auction design aligns investment incentives using a combination of down payment and royalty payment, but inefficiently delays sale and investment. I also characterize informal negotiations as timing and signaling games in which bidders can initiate an auction and determine the forms of bids. I show that post-auction investments are efficient and bidding equilibria are equivalent to those of cash auctions. However, in this setting, bidders always initiate the informal auctions inefficiently early. In addition, I provide suggestive evidence for model predictions using data from the leasing and exploration of oil and gas tracts, which leads to several ongoing empirical studies. Altogether, these results reconcile theory with several empirical puzzles and imply novel predictions with policy relevance. In Chapter 2, I examine learning as an important source of managerial flexibility and how it naturally induces information asymmetry in decentralized firms. Timing of learning is crucial for investment decisions, and optimal strategies involve sequential thresholds for learning and investing. Incentive contracts are needed for learning and truthful reporting. The inherent agency conflicts alter investment behavior significantly, and are costly to investors and welfare. But contracting on learning restores efficiency with low future uncertainty or sufficient liquidity. Unlike prior studies, the moral hazard of learning accelerates good projects and delays bad projects. Even the best type's investment is distorted, and only when learning is contractible can adverse selection dominate learning.