It is possible to estimate the state of the economy through news stories, according to a study from Yale University professor Bryan Kelly and his team, reports Simon Rodda of Dow Jones.
Rodda writes, “Furthermore, modeling economic activity in terms of news-based topics offers economists a new set of tools for understanding shocks.
“Through a novel approach outlined in the paper, titled The Structure of Economic News, he demonstrates that measures derived from textual analysis ‘accurately track a wide range of economic activity measures and that they have incremental forecasting power for macroeconomic outcomes, above and beyond standard numerical predictors.’
“Kelly applies a topic model to a corpus of over 800,000 news articles published by The Wall Street Journal between 1984 and 2017 and extracted from the Dow Jones News Archive. This topic model summarizes business news as easily interpretable topical themes and quantifies the proportion of news attention allocated to each theme at each point in time. These news attention estimates are then used as inputs into statistical models of numerical economic time series.
“He justifies his interest in news as a source of valuable insight thus: ‘The media sector, as a central information intermediary in society, continually transforms perceptions of economic events into a verbal description that we call ‘news’… news text is a mirror of the prevailing economic issues that are important to both news consumers and producers. News text is a one-stop-shop for simultaneously understanding many facets of the state of the economy and how they interact with each other. It uses the richness of narrative, derived from human understanding and interpretation of complex contexts.'”
Read more here.
