Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We decompose the textual data in a Norwegian business newspaper into news topics and investigate their role in predicting and explaining economic fluctuations. Separate full- and out-of-sample experiments show that many topics have predictive power for key economic variables, including asset prices. Unexpected innovations to an aggregated news index, derived as a weighted average of the topics with the highest predictive scores, lead to persistent economic fluctuations, and are especially associated with financial markets, credit and borrowing. Unexpected innovations to asset prices, orthogonal to news shocks and labeled as noise, have only temporary positive effects, in line with economic theory.