Dynamic trade off theory suggests that firms let their leverage ratios vary within an optimal range. I develop an empirical model that estimates how the determinants of capital structure affect the two boundaries that define firms' optimal leverage ranges. Empirical evidence supports the predictions of dynamic trade off theory. Volatility increases the optimal leverage range. Profitability and interest rates reduce the leverage range. Consistent with agency theory, assets in place firms respond sooner to decreases in leverage than do growth firms. I also examine post-adjustment leverage ratios and find that firms under-adjust on average. This suggests proportional adjustment costs play an important role in determining the size of capital structure adjustments.