Tweeting inflation: real-time measures of inflation perception in Colombia

Abstract

This study follows recent developments in the literature using Twitter, now known as X, to measure inflation perceptions. By applying machine learning techniques, we implement two real-time indicators of inflation perception for Colombia and show that both exhibit a dynamic similar to that of inflation and inflation expectations between January 2015 and March 2023. This suggests that our indicators are closely related to the underlying factors that drive inflation. Additionally, we find that our indicators improve the forecast accuracy of inflation and inflation expectations up to six months. Overall, this approach provides a valuable instrument for gauging public sentiment towards inflation, and complements the traditional indicators used in the inflation-targeting framework.