Estimating the Lagging Error in Real Estate Price Indices
提出一个状态空间模型，联合估计房地产价格指数中的潜在真实回报和滞后误差，并利用外生变量和滞后误差来源提高估计精度。去除滞后误差后，指数回报更具信息性。
Real estate indices based on appraisals or sale prices of properties are known for their slow response to market news. These indices can therefore be represented (in logarithm) as the sum of a latent “true” price index and a lagging error. We show that the latent appreciation return and the lagging error can be jointly estimated in a state–space model, which has two key features. First, it employs exogenous variables known to predict asset returns to predict the latent appreciation return. Second, it incorporates known sources of the lagging error, such as the partial adjustment in observed index to the latent appreciation return and the seasonality in reappraisal quality. We find that, after the estimated lagging errors are removed, the appraisal–based National Council of Real Estate Investment Fiduciaries returns become more informative and hence exhibit (i) greater variance, (ii) weaker auto correlation, (iii) higher correlation with the returns of the securitized real estate and (iv) more timely response to market news.