In this paper I apply the Log Periodic Power Law Model (LPPL) to identify bubbles in housing markets. The model overcomes the difficulties that conventional bubble tests face in finding accurate estimates for fundamental house prices. The LPPL Model identifies bubbles by describing the stochastic path of bubbles instead of trying to estimate some fundamental house prices. The model draws on the theory of self-organized systems and methods of statistical physics. I apply the LPPL Model to data on 19 metropolitan housing markets in the US over the last 24 years. My results indicate that the model can reliably identify housing bubbles at early stages. The turning point of a bubble, i.e. the crash, can be predicted with the model within a reasonable time span.