Forecasts of fertility and mortality rates are two of the components of population projections. Because the data consists of age-specific rates, the production of an age distribution curve leads to a high-dimensional problem, where there may be more ages in the data than there are observed years. This paper focuses on forecasting the total rates of fertility and mortality for a set of gender and race/ethnic subgroups, as well as the relative age-specific rates (or age distribution curves) for each. The forecasting of total rates is relatively straightforward, with comparisons done using time series models with and without a drift parameter. For age-specific rates, principal components analysis is used to reduce the dimensionality of the data. The procedure for carrying out the principal components as well as all time series modeling of the principal component series is discussed; included is some discussion of how to choose the number of principal components to retain. Results are produced using fertility and mortality data dating from 1989 to 2009. For total rates, it is determined that a model without drift produces more tenable forecasts in comparison to the occasionally implausible results from the model with drift. Some discussion of ways to improve the procedure for the future are provided.