Firms raise money from banks and the bond market. Banks sell loans in a secondary market to recycle their funds or to trade on private information. Liquidity in the loan market depends on the relative likelihood of each motive for trade and affects firms' optimal financial structure. The endogenous degree of liquidity is not always socially optimal: There is excessive trade in highly rated names, and insufficient liquidity in riskier bonds. We provide testable implications for prices and quantities in primary and secondary loan markets, and bond markets. Further, we posit that risk‐based capital requirements may be socially desirable.

We examine how the payment processing role of banks affects their lending activity. In our model, banks operate in separate zones, and issue claims to entrepreneurs who purchase some inputs outside their own zone. Settling bank claims across zones incurs a cost. In equilibrium, a liquidity externality arises when zones are sufficiently different in their outsourcing propensities—a bank may restrict its own lending because it needs to hold liquidity against claims issued by another bank. Our work highlights that the disparate motives for interbank borrowing (investing in productive projects and managing liquidity) can have different effects on efficiency.

In our parsimonious general‐equilibrium model of banking and asset pricing, intermediaries have the expertise to monitor and reallocate capital. We study financial development, intraeconomy capital flows, the size of the banking sector, the value of intermediation, expected market returns, and the risk of bank crashes. Asset pricing implications include: a market's dividend yield is related to its financial flexibility, and capital flows should be important in explaining expected returns and the risk of bank crashes. Our predictions are broadly consistent with the aggregate behavior of U.S. capital markets since 1950.

We model a dynamic limit order market as a stochastic sequential game with rational traders. Since the model is analytically intractable, we provide an algorithm based on Pakes and McGuire (2001) to find a stationary Markov‐perfect equilibrium. We then generate artificial time series and perform comparative dynamics. Conditional on a transaction, the midpoint of the quoted prices is not a good proxy for the true value. Further, transaction costs paid by market order submitters are negative on average, and negatively correlated with the effective spread. Reducing the tick size is not Pareto improving but increases total investor surplus.