Residential length of stay in the context of a distributed-lag model

Residential length of stay, or alternatively the turnover of familiesin a residential area, is expressed in terms of a distributed-lag model. The parameters of the rational form of the distributed-lag model are estimated from cross-sectional data via a nonlinear maximum-likelihood procedure reported by Nelson and Schwert (1974). The empirical application is based upon data provided by Marshall (1971), and the accuracy of the distributed-lag model in terms of predicting the number of families leaving is compared with the results from Marshall's fitted probability distributions. The usefulness of the distributed-lag model for simulation and control analyses is noted.