Central banks are increasingly using public communication to explain monetary policy decisions and help anchor inflation expectations. This column analyses news articles on Federal Open Market Committee decisions between 1994 and 2023 to examine whether the media accurately covers the Committee’s communication, and whether this coverage influences households’ inflation expectations. The authors find that the media reflects the Committee’s monetary policy sentiment accurately in general, but the accuracy does vary over time. They also find that media coverage significantly influences households’ inflation expectations, especially during periods of high and volatile inflation.
Over the past few decades, central banks have increasingly used public communication to explain monetary policy decisions and help anchor inflation expectations. A rich literature has shown that central bank communication can strongly affect financial markets (McMahon and Hansen 2016, Schrimpf et al. 2018, Pham et al. 2021, Gardner et al. 2022).
However, much less is known about the impact on the broader public. Survey evidence indicates that households acquire information about monetary policy primarily from traditional media (Gardt et al. 2021, Conrad et al. 2022). Hence, a necessary condition for central banks to effectively influence the public is that the media should cover their communication accurately.
In recent work (De Fiore et al. 2024), we leverage improvements in large language models (LLMs) to examine the degree of consistency between the Federal Open Market Committee (FOMC)’s communication and media coverage. More specifically, we use GPT-4 to extract the sentiment regarding the stance of monetary policy – whether dovish, neutral, or hawkish – expressed in Federal Reserve communication and in media coverage across eight major news outlets. The analysis covers 224 FOMC meetings and 14,000 articles between 1994 and 2023.
The media generally covers FOMC communication accurately but with some variation
We find, in general, a high degree of consistency between the monetary policy sentiment expressed in FOMC communication and media coverage, but with considerable variation over time. Figure 1 shows the strength of the FOMC communication pass-through, measured by regressing media sentiment on FOMC sentiment in rolling windows of 70 FOMC meetings. The strength of the communication pass-through declined considerably after the 2008 financial crisis, when policy rates became constrained by the zero lower bound. During this period, the analysis suggests that the Fed faced communication challenges in explaining how unconventional monetary policy tools (e.g. quantitative easing and forward guidance) would provide accommodation.
The communication pass-through began to improve in 2011 with the introduction of the post-FOMC press conferences. These events provide journalists with the opportunity to clarify their understanding of FOMC communication by directly posing questions to the Fed chair. Our findings show that the Fed chair’s answers have become the most effective tool in shaping media coverage. However, the analysis also provides a cautionary message about these tools. The sentiment expressed in the press conference answers is at times not fully aligned with the message conveyed in FOMC written communication. This discrepancy may partly stem from the Fed chair’s efforts to correct misinterpretations of the FOMC decisions that arise from journalists’ questions. But it may also reflect the inherent communication challenges of addressing a wide range of questions under difficult circumstances. The press conferences are thus high-stakes events that can strongly shape media coverage but may also provide a message not entirely consistent with written FOMC communication.
Figure 1 Pass-through of central bank communication to the media coverage
Notes: Pass-through coefficient estimated over overlapping windows including 70 FOMC meetings. Each point on the x-axis corresponds to the date of the FOMC occurring in the middle of the rolling window. The grey area indicates the 90 percent confidence interval.
We also document that the pass-through of FOMC communication to media coverage is significantly weaker in the first few months of a Fed chair’s tenure. This underscores the challenges faced by new chairs in introducing their communication style and establishing credibility with the media.
The media coverage of monetary policy influences households’ inflation expectations
Besides assessing the strength of the FOMC communication pass-through to the media, we also investigate whether the monetary policy sentiment portrayed in the media influences households’ inflation expectations. To this end, we use data from the NY Fed’s Survey of Consumer Expectations, which is available starting in 2013. The analysis examines whether changes in media sentiment around FOMC meetings lead to changes in households’ inflation expectations collected in the days before and after the FOMC meeting. To assess robustness, the analysis is conducted using different time windows, ranging from 3 to 12 days.
Figure 2 shows that when media sentiment becomes more hawkish, households tend to reduce their medium-term inflation expectations. The quantitative effects are sizeable. Panel (a) reports the estimates based on the entire sample of analysis, from 2013 to 2023. A one standard deviation hawkish shift in media sentiment tends to lower three-year-ahead inflation expectations by about 0.18 percentage points. These results provide encouraging evidence about central banks’ ability to affect households’ inflation expectations via the media.
It is also interesting to note that the impact of media sentiment on inflation expectations has been considerably stronger in the post-pandemic period, as shown in panel (b). During this period, the post-pandemic surge in inflation and central banks’ explicit efforts to manage inflation expectations through public communications have likely heightened households’ attention to monetary policy news, thereby amplifying its influence on inflation expectations. This suggests that media coverage of monetary policy tends to have much stronger effects on households’ inflation expectations during times when inflation is high and volatile.
Figure 2 The impact of media sentiment on households’ 3-year-ahead inflation expectations
Notes: The figure reports point estimates and 90% confidence intervals of how changes in media sentiment impact 3-year-ahead inflation expectations over different daily windows around FOMC meetings. The left panel considers the full sample of analysis, from 2013 to 2023. The right panel covers the period from 2020 onwards.
We also show that the results are robust to controlling for the monetary policy sentiment directly extracted from the FOMC communication. In fact, the FOMC sentiment has no effect on households’ expectations. This result is consistent with the premise of the analysis, namely, that households do not directly read FOMC statements or listen to the press conferences but learn about monetary policy from the media.
Conclusion
The analysis provides encouraging evidence about central banks’ ability to reach the broader public via the media. FOMC communication is generally accurately portrayed in the media, with post-FOMC press conferences playing a critical role in influencing news coverage. In turn, the monetary policy sentiment expressed in the media tends to influence households’ inflation expectations, with a hawkish tone lowering expectations. These effects have been particularly pronounced during the post-pandemic inflation surge, when households were arguably more attentive to monetary policy.
Source : VOXeu
