Andrea Guerrieri D'Amati

Email: Andrea dot Guerrieri-D-Amati dot 1 at warwick dot ac dot uk

Office: S1.115

I am a fourth year PhD candidate under the supervision of Roger Farmer, Christine Braun, and Leonardo Melosi.

Research Interests:

Working papers:

Embracing the Future: Tense Patterns and Forward-looking Central Bank CommunicationLink opens in a new window (with Gavin Hassal).
- "This paper studies how forward-looking language in the Federal Open Market Committee (FOMC) minutes affects market expectations of future interest rates. We analyse the text of the FOMC minutes from 1997 to 2023 with structural topic modelling combined with LLM-based tone and tense analysis. We estimate market reactions in an event study that exploits the fact that the release of the minutes involves no policy change, ensuring any market response reflects pure expectation revisions. We show that forward-looking information about certain topics has systematically moved private sector expectations of future interest rates. In particular, hawkish forward-looking inflation language raises 2-, 5- and 10-year Treasury yields. We interpret these findings through a model where the private sector does not observe the central bank's responsiveness to its inflation outlook, and learns about it via a signal extraction problem. We argue that communication effectiveness depends not only on what topics are discussed but on how they are temporally framed."

Work in progress:

Fear, Anger, and the Price of Risk.
- "This paper asks whether emotions in economic news affect the market price of risk. I combine 9.83 million UK newspaper articles with FTSE 100 returns from 2003 to 2022. I measure daily fear and anger using dictionary word shares and LLM-classified article shares. A portfolio-choice model motivates a conditional mean--variance ratio as a proxy for the price-of-risk. To address common shocks, I instrument economic-news emotions with matching emotions in non-economic and sports news and relate emotions and controls at \(t\) to the price-of-risk state at \(t+1\). A one-standard-deviation increase in instrument-predicted dictionary anger raises the proxy by 34.0 per cent of its mean with non-economic-news instruments and 31.3 per cent with sports-news instruments, implying lower risk appetite. Fear changes sign across measures: dictionary fear is negative, whereas LLM fear is positive."

New Neighbours: The Effects of Corporate Relocation on Local Labour Market Expectations (with Christine Braun).

I have had the pleasure of (co) organising the following seminar series and conferences: