We develop a model of algorithmic trading in which a better-informed algorithm competes with human traders who choose whether to acquire costly information and participate in the market. We compare a model-free algorithm, which learns how to trade based on realized profits alone, with a model-based algorithm that learns from a model of its price impact. We show that model-based learning induces more aggressive trading,...

We study pricing dynamics and risk-sharing in a market with rational investors and a Q-learning trader. The Q-learner's trading generates a feedback loop in prices: their demand for the risky security depends on their perceived benefit from trading, which in turn, depends on realized returns. We show that this loop generates state-dependent stochastic volatility, predictable returns, and novel price dynamics which depend on the mass...

Common wisdom suggests that greater climate-risk disclosure improves welfare. We show that this need not be true. A firm endogenously chooses whether to adopt a costly green project to maximize its stock price. When disclosure reveals the project's climate-risk exposure (``greenness"), mandatory transparency reduces green adoption and can lower investor welfare. Costly voluntary disclosure can improve outcomes relative to both mandatory disclosure and no disclosure...