This study explores the physical mechanisms by which the duration of strong ground motion influences structural response. While a number of previous studies have found that ground motion duration influences only cumulative damage indices, and not peak structural deformations, a few recent studies that employed realistic, deteriorating structural models were able to demonstrate the effect of duration on peak deformations and structural collapse capacity. These recent studies were, however, empirical in nature and did not fully explore the reasons behind the observed effects of duration. Many of the previous studies qualitatively attributed the effects to the cyclic deterioration of strength and stiffness of the structural components, which represents just one mechanism by which duration exerts its influence. In contrast, the present study shows that the gradual ratcheting of drifts, accentuated by the destabilizing P − ∆ effect, is an equally important mechanism by which duration influences structural response. The relative contributions of the two mechanisms—cyclic deterioration and ratcheting—to the observed influence of duration on the collapse capacity of a five-story steel moment frame building, are quantified by conducting incremental dynamic analysis (IDA) using spectrally equivalent sets of long and short duration ground motions. The use of spectrally equivalent ground motions allows controlling for the effect of response spectral shape. A response parameter called the ratcheting interval is defined and used to explain the larger potential for a long duration ground motion to cause structural collapse, when compared to a spectrally equivalent short duration ground motion scaled to the same intensity level. These findings shed light on the interaction between structural model characteristics and the observed influence of ground motion duration on structural response. In addition, they highlight the importance of using models that capture both cyclic deterioration and the P−∆ effect to reliably account for the effect of ground motion duration when assessing structural collapse risk.