Conventional earthquake resistant design methods are mainly set up to estimate the maximum values of response parameters, such as displacements and forces. As a result, most of the finite-element modeling approaches and tools have been verified for estimating these maximum values. However, in the last decade there has been an increasing interest in the direct consideration of post-earthquake residual displacements in the design and assessment processes. In this study, it is shown that the residual displacement estimates computed using nonlinear dynamic finite-element analyses can be strongly influenced by the adopted modeling approach and the simplifications introduced. Nonlinear response of a representative reinforced concrete column to static cyclic and dynamic loads is simulated using various implementations of lumped plasticity and fiber models. Comparative evaluation of the results shows that significantly different residual displacements are computed using different implementations of the same models. On the other hand, the maximum displacements are found to be less sensitive to the differences among implementations. Results indicate that the assumptions related to temporal and spatial discretization of the system has an important influence on the computed residual displacements. Copyright © (2006) by Earthquake Engineering Research Institute.