Persistent performance in venture capital is routinely interpreted as evidence for skill. We present a dynamic model of delegated investment with endogenous fund heterogeneity and deal flow, which generates performance persistence without skill differences and predicts mean reversion in long-term performance. Investors working with multiple funds use contingent payments and tiered contracts to induce proper project nurturing and managerial effort. Successful funds receive continuation contracts that tolerate investment failure and encourage innovation, and subsequently finance entrepreneurs through a path-dependent assortative matching favoring incumbents. Recent empirical findings corroborate the model’s general implications, and the economic mechanisms are robust to short-term contracting, endogenous bargaining, and double moral hazard issues.

We develop a dynamic model of a platform economy where tokens serve as a means of payment among platform users and are issued to finance investment in platform productivity. Tokens are optimally rewarded to platform owners when token supply (normalized by productivity) is low and burnt to boost franchise value when the normalized supply is high. Although token price is determined in a liquid market, the platform’s financial constraint generates an endogenous token issuance cost that causes underinvestment through the conflict of interest between insiders (owners) and outsiders (users). Blockchain technology mitigates underinvestment by addressing the owners’ time inconsistency problem.