Shipping sentiment and the dry bulk shipping freight market: new evidence from newspaper coverage

Abstract

In this paper, a shipping sentiment index in the dry bulk market is constructed using computational text analysis from shipping news archives. Our news corpus consists of 11,296 dry bulk news headlines from two major shipping news websites from January 2014 to November 2020. The constructed monthly index does track closely with major dry bulk freight indices. Then, we investigate how the freight market responds, if at all, to sentiment shock and whether news sentiment helps predict freight rates. Specifically, impulse response analyses using vector autoregression (VAR) models are employed to examine the effect of sentiment shocks on freight indices. The predictive power of the constructed shipping sentiment index for future dry bulk freight rates is further investigated using a least absolute shrinkage and selection operation (LASSO) regression. To examine potential asymmetric relationship between dry bulk freight rates and news sentiment, copula models are then utilized. Results show that the constructed news sentiment index is a significant predictor for the future freight rates, especially in the Capesize market. The effect of sentiment shocks is relatively transitory and varies across different dry bulk sub-segments. Last but not least, the impact of news sentiment on the dry bulk freight rates is asymmetric and tends to be higher when the sentiment is extremely positive, and such tail dependence is more obvious for the larger Capesize segment. This study provides significant implications and decision support for shipping players.