I explore the timing of and effects of the U.S. financial crisis of the 1930s in a regime switching framework. Estimated conditional probabilities over the state of the financial sector suggest that a prolonged period of crisis begins not with the 1929 stock market crash, but with the first banking panie of October 1930. These probabilities also suggest that the crisis persists until the introduction of federal deposit insurance in early 1934. Consistent with the view that this financial crisis had real effects, these conditional probabilities contain additional explanatory power for output fluctuations. This is in addition to that provided by the money stock.