Paper sheds light on historic economic sentiment

New research analysing 200 million news pages featured by Financial Times

Newspaper Historic - 1140x346

New research co-authored by London Business School Assistant Professor of Finance Svetlana Bryzgalova and a Ph.D. candidate in Finance Mayukh Mukhopadhyay sheds light on 170 years of economic sentiment.

The National Bureau of Economic Research working paper, featured by the Financial Times, combines text data from 200 million news pages and machine learning techniques to identify both country- and state-level economic sentiment in the U.S.

The authors show that news-based economic sentiment leads GDP growth at the country and state level, employment, consumption, and even monetary policy during recessions. It is distinct from the expert economic growth forecasts and leads its consensus value. 

The paper also shows that the tone of news coverage has become increasingly negative over the last 50 years, for which the researchers adjusted their model.