Can Risks be Good News? Revealing Risk Perception of Real Estate Investors using Machine Learning

Abstract

The SEC mandates firms to inform investors about their assessment of future contingencies in their 10 Ks. However lengthy and complex disclosures – mostly for dozens of firms in an investor’s portfolio – can barely be processed by a human being. To cope with the flood of information, we exploit an unsupervised machine learning algorithm, the Structural Topic Model, to identify and quantify the risk factor topics discussed in the 10-K filings. We apply this algorithm to a US REIT sample between 2005 and 2019 to assess whether the proportion of the disclosure a firm allocates to a specific risk factor provides new information and how Item 1A affects investor risk perception. Our results suggest, that the topic distribution chosen by the REIT’s manager is significantly associated with stock return volatility after the filing submission date. We conclude, that REITs provide previously unknown information in their risk disclosures, approximated by our topic allocation, leading to a market reaction. Furthermore, we investigate whether and how individual topics affect the risk perceptions of investors. We find all three kinds of topics: uninformative topics with no impact, increasing risk perception topics, and decreasing risk-perception topics which is the majority. The predominance of the risk-reducing effect indicates that risks can indeed be interpreted as good news.

Keywords: Real Estate Investment Trust (REIT), risk, text analysis, machine learning, Latent Dirichlet Allocation, Structural Topic Model, 10-K, Item 1A, Item 7A

JEL Classification: C45, C80, G14, G18, K22, K40, M41, M48, R30