Temporal aggregation is a repeat sale index construction methodology that consists of aggregating paired‐transactions in a moving‐average window. In particular, the methodology is used to calculate the popular S&P/Case‐Shiller home price indices. In this article, I focus the insights of the literature on measurement error to demonstrate that temporal aggregation produces idiosyncratic biases in predictive regression slopes. I further estimate a dynamic instrumental variable (IV) panel for the 20 S&P/Case‐Shiller metro areas. The main empirical finding is that temporal aggregation is a short‐lived statistical disturbance that does not explain the homogenous robust persistence of the indices.