Early warning indicator systems for real estate investments: Empirical evidence and some thoughts from the perspective of financial risk management

Abstract

In recent years, early warning indicators and real estate—as an increasingly relevant asset class—have received more and more attention in German insurers’ investment strategies. Therefore, this paper examines the relationship of real estate sentiment data as leading indicators for housing activity and house price indices. However, we focus on empirical evidence from the US due to the quite limited data availability in Germany. The National Association of Home Builders (NAHB) housing market index is used as an indicator for US real estate prices and other variables related to housing activity and the S&P/Case-Shiller 20 city home price index for house prices in the US. In order to test for Granger causality among US house prices and the NAHB sentiment indictor we employ a modified Wald test based on Toda and Yamamoto (1995) examining an augmented vector autoregressive (VAR) model in levels. The results of our empirical investigations do show that there are clear signs for unidirectional Granger causality running from the NAHB housing market index to the S&P/Case-Shiller index. Therefore, the NAHB data seem to be quite helpful predicting US house prices. This empirical finding is of high relevance with regard to the construction of early warning indicator systems for real estate prices.