Publisher:
Rutgers University, Department of Economics, New Brunswick, NJ
Abstract:
This paper investigates sentiment in the US economy from 1920 to 1934 using digitized articles from the Wall St Journal. We derive a monthly sentiment index and use a ten variable vector error correction model to identify sentiment shocks that are orthogonal to fundamentals. We show the timing and strength of these shocks and their resultant effects on the economy using historical decompositions. Intermittent impacts of up to fifteen percent on Industrial Production, ten percent on the S&P 500 and Bank loans and, thirty-seven basis points for the Credit risk spread, suggest a large role for sentiment.