Topic-based uncertainty using news articles in Dagens Næringsliv

I create measures of uncertainty by first classifying news articles according to theme, and then quantifying uncertainty by the count of uncertainty terms within the different types of news. The uncertainty measures are available at a daily frequency and capture well-known events linked to uncertainty both at an aggregate and a category-specific level. The paper can be found here.


Selected uncertainty measures (updated August 31. 2020):

Housing uncertainty:

T5 uncertainty


Macroeconomics uncertainty:

T19 uncertainty

Oil price uncertainty:

T33 uncertainty

The figures plots the 200 days backward-looking mean of the count of uncertainty terms within classified news. The news topics can be represented as word clouds and these word clouds for all the 80 news topics can be found here (the page can take some time to load).