Firms often choose between concentrated, renegotiable bank claims and dispersed, arm’s-length market debt. I develop a tractable adverse-selection model in which both financiers can take collateral, but they differ in enforcement and coordination efficiency at default. The single primitive wedge—a higher effective liquidation rate for concentrated creditors—is sufficient to generate coexistence, a sharp cutoff in the bank–bond partition, and distinctive comparative statics. A marginal improvement in bankruptcy/insolvency efficiency or bondholder coordination reallocates issuance toward bonds, compresses loan–bond pricing gaps for safe types, and shifts default incidence and collateral intensity in predictable ways. The welfare decomposition clarifies when strengthening bank enforcement reduces deadweight liquidation losses (by moving marginal types into monitored finance) versus when improving market-side coordination dominates (by accelerating dispersed-creditor resolution). The model delivers stability-relevant predictions for default rates, recovery, covenant stringency, and issuance composition around reforms that move liquidation efficiency on either side, and it provides a disciplined mapping to empirical proxies (recoveries, covenant strength, creditor dispersion).