Spillovers between US real estate and financial assets in time and frequency domains

摘要

Purpose - Assessing the strength and time variation of spillovers between returns on residential real estate, real estate investment trusts (REITs), stocks and bonds in the United States. Spillovers reduce the benefits of portfolio diversification, especially in crisis times, when asset returns tend to be more correlated. Design/methodology/approach - The Diebold-Yilmaz approach in the time domain and the Barunik-Krehlik methodology in the frequency domain are used. The latter allows distinguishing spillovers generating only short-lived volatility from those with a more persistent effect. Findings - On average, spillovers between housing, stock and bond returns are relatively modest and shocks to stock and bond markets affect housing returns more than the other way round, even though with variations over time. Spillovers in both directions are much stronger between REITs and stocks than between REITs and housing. Shocks originating in the housing market are most persistent, particularly in the aftermath of the subprime crisis. Practical implications - Housing provides a hedge against volatility in financial (including REITs) markets. However, hedging strategies involving housing need to take into account potential tail events such as the GFC and the investment horizon. Originality/value - To the best of the knowledge of the authors, this paper is the first to apply the Barunfk-Krehlik methodology to real estate price spillovers. Although the Diebold-Yilmaz methodology has been used in several studies on spillovers between residential real estate and financial asset returns, this paper covers a new set of variables and time span.