Is the Housing Market in the United States Really Weakly-Efficient?

ABSTRACT

We analyse the directional predictability of a daily dataset of aggregate and regional (10 major metropolitan cities) housing markets of the United States using the quantilogram – a model-free procedure. We overwhelmingly reject the weak-form of the efficient market hypothesis (EMH), which has been derived thus far by the extant literature based on unit root tests and long-memory models.

KEYWORDS:

• Correlogram
• dependence
• quantiles
• efficiency
• housing markets
• US

JEL CLASSIFICATION:

• C22
• R31

Acknowledgments

We would like to thank two anonymous referees for many helpful comments. However, any remaining errors are solely ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The data is downloadable from http://qed.econ.queensu.ca/jae/datasets/bollerslev001/.

² In fact, Wang (2014) discusses in detail that the daily house price indices are indeed non-stationary in levels based on unit root tests, and hence, the housing market can be again be concluded to be weakly efficient. In addition note that, to ensure that our results are not driven by the daily frequency of the data, we also applied the quantilogram on the log-returns of the monthly version of the index (derived from the FRED database of the Federal Reserve Bank of St. Louis), and reached similar conclusions, i.e. the null of no-predictability was strongly rejected. Complete details of these results are available upon request from the authors.

³ We also conducted the rolling quantiles-based portmanteau test for the 10 MSAs, and found the pattern of rejection of the null hypothesis to be similar as that for the overall housing returns. We have suppressed these results in the main text to save space; however, complete details are available upon request from the authors.