Quantitative structure and sentiment of news articles on housing policies

Housing policies greatly affect real estate markets, and the media quickly responds to the tone and consistency of housing policies. In this sense, news articles can be utilized to comprehend the relationship between market participants' reactions and housing policies. Based on daily news articles from three major media outlets covering the South Korean housing market, we employ the controlled growth process model to investigate the quantitative structure of news articles on housing policies and their association with embedded sentiment. Our findings reveal that word frequencies in news articles follow a power-law distribution, and a news article can be considered a semi-structured document in terms of the intermediate-level text cohesion between technical reports and narrative texts. Furthermore, differences in scaling exponents and text cohesion can explain the heterogeneous sentiment patterns of news articles on housing policies (i.e., the first- and second-order effects on sentiments in news articles). This study contributes to the existing literature by providing an extended window for understanding how linguistic patterns in word frequency distributions are interlinked with embedded sentiments. Regulators and policymakers can consider our theoretical framework for ex post policy evaluation, obtaining insights into scheming forward housing policies to relieve real estate markets. Investors can benchmark this analytical procedure in monitoring the housing market's responses to announced policies to adjust their strategies regarding real estate financing in a timely manner.