Monetary Policy in the Age of Social Media: A Twitter-Based Inflation Analysis


We develop a high-frequency inflation index derived from German tweets through sophisti- cated NLP methods. This index aligns closely with realized inflation and consumer inflation expectations, both nationally and regionally, offering enhanced predictive precision over current benchmarks. Notably, it responds to monetary policy shifts, rising post-easing and falling after unexpected tightenings. The influence is particularly pronounced from tweets by private users during recent periods of elevated inflation. Elevated inflation expectations correlate with reduced consumer spending, as gauged from online transaction data, particularly on discretionary goods. Consequently, this Twitter-centric index offers a valuable real-time tool to assess prevailing inflation sentiments.