Abstract

The Consumer Confidence Index (CCI) measures public sentiment about the economy through a probability sample survey considering four questions regarding household finances, economic outlook, and spending plans over the past and coming years. It is calculated as a mean of these responses. The main objective of the study is to nowcast the CCI, estimating the current month’s values faster than traditional survey methods, which usually provide results at the end of the month. We examine the relationship between traditional survey-based indicators and consumer sentiment expressed on social media. Social media expressions are collected from X (Twitter). The sentiment analysis of tweets enables us to create a Social Media Indicator (SMI), offering a distinct advantage in our predictive models. To improve forecast accuracy, we include Google Trends data, providing additional insights into consumer search behaviour and related trends in economic confidence. The study also explores integrating key economic indicators such as inflation rate, income statistics, and unemployment. In addition to nowcasting the CCI at the national level, we use county-specific auxiliary information including Google Trends data to improve the direct CCI estimates at the county level through small-area estimation techniques. This approach allows us to nowcast the CCI at the county level as well.