Essays on Informational Efficiency in Real Estate Markets

In order to contribute to a better understanding of real estate markets with respect to their informational efficiency, this thesis seeks to shed light on two problem areas: First, the effect of noise on the market, and second, the effect of information withheld from the market. The central objective of the first part of this thesis (Paper 1 and 2) is to develop a sentiment measure for the real estate market which overcomes fundamental limitations of existing ones. The approach utilizes an artificial neural network to classify news articles on the US real estate market. The classification process does not rely on manual classification, but instead uses a large text sample with distinct sentiment-polarity, so as to train the network via a distant supervision approach. The developed sentiment index is used to make predictions on the direct commercial real estate market in the US in terms of market returns (Paper 1) and trading liquidity (Paper 2). The results suggest a significantly positive relationship of the sentiment index to future returns of the US direct real estate market. Application of a Markov-switching model exhibits a varying influence of sentiment on returns during up- and down-market periods, with a stronger relationship during down markets. The sentiment indicator is furthermore tested in a logit forecasting framework. In this context, it exhibits predictive potential, however struggles with short-term market fluctuations. Regarding market liquidity, the sentiment index is positively related to contemporary and future price impact of trading as well as trading volume, with a more pronounced effect on the price impact measure of liquidity. The second part of the thesis (Paper 3) deals with the converse problem of information which is not available publicly. More precisely, the disposition of executives of property-developing Real Estate Investment Trusts (REITs) to hoard bad news is analyzed. In the existing literature, bad news hoarding is widely considered to be the origin of stock price crashes of listed companies. Furthermore, the entitlement of equity REITs to develop their own properties is an actively debated matter in both real estate academia and practice. The paper finds that a higher share of property development activity leads to a decrease in future stock price crash risk for the examined sample of US equity REITs. The empirical evidence indicates that executives of property-developing REITs employ a long-term view on company performance and refrain from short-termism practices like bad news hoarding