Using administrative records to study mobility: the case of the housing assistance supply experiment
A new, efficient statistical technique, the length-of-stay procedure, analyzes residential mobility histories. It analyzes mobility not as a move's simple occurrence or nonoccurrence during some arbitrary period of time, but through limited data on the interval between moves. Typically, our knowledge of this interval is "censored" by short administrative or survey records. The paper applies the procedure to short-period data from RAND's Housing Assistance Supply Experiment. It shows the major role played by eligibility requirements and also the duration-of-residence effect--the tendency for continued residence to change future mobility rates.