2023-07-05-美联储-联邦公开市场委员会运作目标的使用_85年和计数_28页_2mb

The paper analyzes the Federal Open Market Committee's (FOMC) use of operational targets over 85 years, from 1936 to 2021, using natural language processing (NLP) techniques applied to FOMC meeting minutes. Key findings include peaks in operational target discussions following significant events like the 1951 Treasury-Fed Accord, the 1979 anti-inflation efforts, and the 2007-2009 financial crisis. Multiple operational targets, often rate and quantity pairs, were commonly discussed, with a shift toward single-rate targets starting in the 1990s. Operational regimes identified through NLP methods, such as TF-IDF measures, align with historical periods and show breaks after major economic downturns. Macroeconomic conditions, including improving economic indicators, correlate with regime shifts. NLP complements traditional research by revealing trends in operational target usage, highlighting the FOMC's evolution in policy implementation, and noting factors like the effective lower bound and asset purchases post-crisis.

For instance, regimes display high volatility pre-1960 but stabilize later, with the Great Moderation contributing to operational calmafter 1980. NLP-based regime identification maps to historical assessments and shows distinct clusters from the 1970s monetary aggregates era to the 2020s fed funds dominance. Economometric analyses reveal that factors like moneysupply and external accounts influence regime changes, while economic conditions such as declining unemployment or rising GDP growth trigger shifts. The paper emphasizes NLP's utility in uncovering policy trends without replacing detailed historical work, offering new insights for future research.