Assessing Consumer Confidence from Online Sources: An Abstract

Abstract

Consumer confidence indices are reliable indicators of the way consumers plan to spend their money and of the state or direction of the economy (Starr 2012; Toussaint-Comeau and McGranahan 2006; Wilcox 2007). In this context, the possibility of analyzing big data from social networks in order to assess consumer sentiment, behavior, and trends can have a significant impact on social sciences and economics and can help in nowcasting and forecasting not only at macrolevel but also for companies in creating their marketing strategies. However, not a lot has been done from this point of view in relating new ways of assessing consumer confidence and the relationship of consumer confidence with the stock market (Andruszkiewicz et al. 2013; Barsky and Sims 2012). In this exploratory study, we use the framework of “animal spirits” (Akerlof and Shiller 2009) to explore new ways of measuring consumer confidence through Google search popularity and establish its relationship with stock market fluctuations. Given the importance of consumer confidence indices and of consumer sentiment for markets and the economy, the objective of this study is to formulate and analyze indicators from social media consumer communications and online searches that can measure consumer sentiment, including the five dimensions of the Keynesian “animal spirit.”

The main analysis focused on assessing whether consumer searches of different keywords in Google can be used to estimate consumer sentiment in a different manner than the three major indices calculated based on surveying consumers. The two main contributions of this exploratory study show that consumer sentiment can be measured through online data mining and can produce even more comprehensive and complex consumer confidence measures and these measures can be used to evaluate and anticipate stock market evolutions.

Overall, the most important result and conclusion of this study is the fact that the keywords selected to represent the Keynesian five dimensions and their search indices in Google by consumers from the United States are significantly correlated to the three main consumer confidence indices used by the American market and researchers. From this point of view, we demonstrated that a fast alternative to assessing consumers’ confidence and willingness to spend can be performed by using the data provided by Google Trends.