Reducing the Error in Monthly Housing Starts Estimates

This paper describes a method for developing estimates of total monthly housing starts that are more accurate than the published Census Bureau figures. The technique makes use of the facts that (1) estimated building permit issuance is subject to far less sampling error than is the starts estimate and (2) permit issuance and starts bear a strong contemporaneous correlation. The conclusion is that monthly housing starts and monthly building permit issuance should be assigned nearly equal weights in developing an improved estimate of total housing starts.