We estimate a continuous-time model with stochastic volatility and dynamic crash probability for the S&P 500 index and find that market illiquidity dominates other factors in explaining the stock market crash risk. While the crash probability is time-varying, its dynamic depends only weakly on return variance once we include market illiquidity as an economic variable in the model. This finding suggests that the relationship between variance and jump risk found in the literature is largely due to their common exposure to market liquidity risk. Our study highlights the importance of equity market frictions in index return dynamics and explains why prior studies find that crash risk increases with market uncertainty level.