This paper introduces an improved procedure for estimating capital asset price indexes. We jointly estimate conventional hedonic and repeat sales models via maximum-likelihood procedures, thereby taking advantage of the unique features of the individual models and using all the data that are available. Our model captures depreciation within the repeat sales model and accounts for serial correlation in hedonic data. The improvement in precision obtained by estimating the joint model is illustrated by smaller standard errors and narrower interval estimates for the resulting price indexes. We also carry out a simulation experiment that shows estimation errors significantly smaller using the joint estimation technique than either of the individual models or the GLS estimator of Case and Quigley (1991).