Is there a role for sentiment in describing economic behavior? A new paper by Harold James and co-authors, Ali Kabiri, John Landon-Lane, David Tuckett, and Rickard Nyman, uses digitized articles from The Wall Street Journal (WSJ) from 1920 to1934 to investigate how broad-based shifts in sentiments could have influenced economic outcomes in the US before and after the Great Crash of 1929.

Contemporary market commentators including the economist John Maynard Keynes cited both the exuberance of the roaring twenties and the pessimism in the years after the crash as important factors governing financial and economic behavior at the time. From a research perspective, the problem is that while sentiment is quickly felt and often shared, it is hard to measure and quantify. To measure sentiment, the authors assembled a lexicon of 300 words indicating positive (approach) or negative (avoidance) sentiment and constructed a monthly 'sentiment index' based on how frequently the words appeared in the WSJ. The authors use a vector error correction model to analyze how the timing and strength of shifts in sentiment (or sentiment shocks) affected the economy.  Read the paper...