Public policymakers and program administrators often face decisions that impact the retirement incomes of individuals. An important question that these decision-makers may wish to address concerns the distributional impacts of the programmatic changes under consideration. Who (what population groups) would gain income and how much? Who would be unaffected? Who would lose and by how much? The question that this paper investigates is the extent to which computer models and associated policy analysis capability are available to provide decisionmakers with this kind of information. Specifically, the paper reviews a class of models that may be designated as retirement income policy models. The review is limited to existing models of programs within the U.S. retirement income system. The paper presents a general discussion of the components of such models. It describes features of models that are in the current portfolio and assesses components of those models. It furthermore presents examples of how the existing models might analyze two policy initiatives changing the "bend points" of the Social Security old-age and survivors insurance benefits and increasing the maximum allowable contributions to IRAs. The paper finds that the current capacity is not great, nor is it fully utilized. However, a foundation exists to be able to examine distributional impacts of retirement policy changes. Three concerns that need to be addressed now to assure that this foundation will be useful in the future are (1) assessing whether the existing modeling capacity is substantive enough and flexible enough to gear up in a short amount of time, (2) assessing whether decisionmakers will have adequate resources to support the necessary development and refinement of models, and (3) assuring that appropriate recent data sources are available to support the models.