The Timeline Estimation of Bubbles: The Case of Real Estate

We develop and then apply a new recursive regression methodology with a two‐direction searching process for initialization. The methodology improves the reliability of existing models when estimating a bubble's timeline. We apply our proposed methodology to estimate bubbles in U.S. home prices as well as in simulated scenarios. Our results confirm the improvement in reliability of the proposed methodology in obtaining consistent estimators with varying samples. Moreover, we verify the presence of bubbles in the U.S. aggregate data and seven of the eight cities in our study prior to the subprime crisis and find evidence of the bubble's reemergence since October 2013.