We examine the quantitative importance of financial market shocks in accounting for business cycle fluctuations. We emphasize the role financial markets play in reallocating funds from cash-rich, low productivity firms to cash-poor, high productivity firms. Using evidence on financial flows at the firm level, we find that for publicly traded firms (in Compustat), almost all investment is financed internally. However, using an alternative data source (Amadeus), we find that most investment by privately held firms is financed through borrowing. Motivated by these observations, we build a quantitative model featuring publicly and privately held firms that face collateral constraints and idiosyncratic risk over productivity as well as non-financial linkages. In our calibrated model, we find that a shock to the collateral constraints which generates a one standard deviation decline in the debt-to-asset ratio leads to a 0.5% decline in aggregate output on impact, roughly comparable to the effect of a one standard deviation shock to aggregate productivity in a standard real business cycle model. In this sense, we find that disturbances in financial markets are a promising source of business cycle fluctuations when non-financial linkages across firms are sufficiently strong. * We are indebted to V.V. Chari and Larry Jones for valuable advice. We would also like to thank