Essays on Sentiment Analysis through Textual Analysis in Real Estate Markets [book]

Marcel Lang
2018
What is the current state of research on text-based sentiment analysis for real estate markets? Can sentiment expressed in real-estate-related news headlines be captured via a machinelearning approach? Does text-based sentiment, captured via a supervised machine learning approach, predict returns from the securitized and direct commercial real estate market? What are the intertemporal linkages between news-based sentiment and the securitized and direct U.S. real estate markets? Do the
more » ... sentiment measures provide additional information beyond that provided by established sentiment indicators? Are market participants influenced more strongly by pessimism rather than optimism expressed in news? Hence, is there empirical evidence of a negativity bias? Inspired by the findings from the first study, this paper is the first to characterize the bi-directional relationship between news-based sentiment and the performance of U.S. direct commercial real estate from 2001 to 2016. The direct commercial real estate market is expected to be less efficient than the securitized and/or public market, and hence more susceptible to the impact of sentiment. Using the dictionary-based approach, optimism and pessimism expressed in approximately 65,000 real-estate-related news abstracts of The Wall Street Journal is quantified and aggregated on a quarterly basis.
doi:10.5283/epub.37597 fatcat:eb2ytdb5tbd2ra52iqhdxn6xbm