Explaining US consumer behavior and expectations with news sentiment

We examine the information content of a newly created news sentiment index from over 300,000 articles from some of the most widely read newspapers in the US to explain changes in the University of Michigan Index of Consumer Sentiment from 1995 to 2009. Using ARMA-models, we show that consumer sentiment is influenced by news sentiment and other variables, such as prices, income and interest rates. While there exists a statistically significant relationship between news sentiment and private consumption, the consumption behavior of private households can best be explained by consumer sentiment combined with changes in personal income and consumer prices. We add news sentiment to the causality chain before consumer sentiment and private consumption.