Making text count: Economic forecasting using newspaper text

This paper examines several ways to extract timely economic signals from newspaper text and shows that such information can materially improve forecasts of macroeconomic variables including GDP, inflation unemployment. Our is drawn three popular UK newspapers collectively represent readership in terms political perspective editorial style. Exploiting both unconditionally when conditioning on other relevant information, but the performance latter varies according method used. Incorporating into by combining counts with supervised machine learning delivers highest forecast improvements relative existing text-based methods. These are most pronounced during periods stress when, arguably, matter most.