This paper considers forecasting regressions of "realized volatility" on a misalignment measure. Results show that this misalignment measure is useful to predict in and out-of-sample stock-market volatility at monthly horizons. The analysis also suggests a threshold effect.