Estimating central bank policy preferences often relies on voting data, but consensus voting can obscure individual stances.
This paper introduces a novel methodology combining topic-based text analysis with scaling techniques to gauge the policy preferences of US Federal Open Market Committee (FOMC) members during 2007-2008.
Data & Methods: Topic Modeling on Speeches
It analyzes FOMC speeches using advanced statistical methods to extract sentiment and focus without depending on voting records.
Key Findings:
* Unemployment Focus: Members representing districts with higher unemployment emphasized this issue more in their public statements.
* Voting Schedule Impact: Contrary to traditional measures, those scheduled to vote later showed distinct differences from early voters in opinion divergence.
Why It Matters
This approach provides a new lens for understanding policy leanings during periods of high consensus or non-voting. Applying it reveals insights that challenge conventional wisdom about measuring central bank preferences.
| A Textual Taylor Rule: Estimating Central Bank Preferences Combining Topic and Scaling Methods was authored by Nicole Rae Baerg and Will Lowe. It was published by Cambridge in PSR&M in 2020. |
