This paper studies the integration of deposit and loan markets, which may be constrained by the geographic dispersion of depositors, borrowers, and banks. This dispersion results in problems of asymmetric information, monitoring and transaction costs, which in turn may prevent deposits from owing from areas of low demand for loans to areas of high demand. We provide systematic evidence on the extent to which deposits and loans are geographically imbalanced, and develop a methodology for investigating the contribution of (i) branch networks, (ii) local market power, and (iii) economies of scope to this imbalance using data at the bank-county- year level from the US banking industry for 1998-2010. Our results are based on the construction of an index which measures the geographic imbalance of deposits and loans, and the estimation of a structural model of bank oligopoly competition for deposits and loans in multiple geographic markets. The estimated model shows that a bank's total deposits have a significant effect on the bank's market shares in loan markets. We also find evidence of significant economies of scope between deposits and loans at the local level. Counterfactual experiments show that multi-state branch networks contribute significantly to the geographic ow of credit but benefit especially larger/richer counties. Local market power has a very substantial negative effect on the ow of credit to smaller/poorer counties.