Job Market Paper
Fedspeak, LLM-Derived Signals, and High-Frequency Trading
This paper studies how financial markets respond to changes in Federal Reserve communication beyond conventional monetary policy surprises. I develop an LLM-based Communication Divergence Signal that measures semantic and tonal shifts across FOMC statements, minutes, and press conferences. Using a high-frequency event-study design with futures data, I show that communication divergence generates sizable and persistent movements in Treasury yields and trading activity across the yield curve. The effects are stronger after 2020, suggesting that markets have become more responsive to differences in how the Fed communicates its policy outlook.