The adequate measurement of real estate risk is of utmost importance for asset management and real estate portfolio management. Most real estate academics agree that volatility is inappropriate for that purpose. However, it is still a favored measure of many practitioners and academics due to its simplicity and because the perfect alternative has yet to be found. We provide plausible reasons for the proposition that volatility should not be used for measuring the risk of real estateóneither within its asset class, nor in a multi-asset environment. They are based on an extensive literature overview, expert interviews, and new empirical evidence from Germany. Furthermore, we discuss whether qualitative risk measures might be more appropriate and provide some requirements for better real estate risk measures. Our paper is an update on earlier work on the same subject. Compared to the first version we improved both the data quality and the methodology. Instead of data from two German real estate portfolios we used IPD data for 939 properties from 1996-2009. We performed several analyses, mainly tests on the normality of real estate returns. Furthermore, by following an approach by Lizieri and Ward (2001), we also fitted alternative theoretical distributions to the observed frequency distributions. The results are very much in line with the previous results: The returns of direct real estate investments in Germany are not normally distributed, in fact, most of the return series follow a logistic distribution, hence, the volatility does not adequately capture the real estate risk.