We extend the literature on crash prediction models in three main ways. First, we explicitly relate crash prediction measures and asset pricing models. Second, we present a statistical significance test for crash prediction models. Finally, we propose a definition and a measure of robustness for these models. We apply our statistical test and measure the robustness of selected model specifications of the Price‐Earnings (P/E) ratio and Bond Stock Earning Yield Differential (BSEYD) measures. This analysis shows that the BSEYD and P/E ratios, were statistically significant robust predictors of corrections on the US equity market over the period 1964 to 2014.