Monetary policy and markets: when tweets anticipate financial volatility
Before a monetary policy decision, predictions and interpretations circulate on social media. After the announcement, the conversation is measured against what the central bank actually communicated. A study based on nearly 230,000 tweets measured how closely the messages posted on social media before and after each announcement resemble the official text, assigning two separate numerical values to pre- and post-announcement similarity. The greater the change between the two values, the stronger the reaction in financial markets tends to be.
The research by Donato Masciandaro, Davide Romelli , and Gaia Rubera analyzes nearly 230,000 tweets related to decisions by the European Central Bank, the US Federal Reserve, and the Bank of England. When, after an announcement, the content of messages changes substantially compared with the preceding conversations, both volatility and the magnitude of changes in stock and government bond prices increase.
This relationship, which is particularly strong for the ECB, emerges mainly during press conferences, rather than when the initial statement is released. Understanding the market reaction, therefore, requires looking not only at the decision itself, but also at the explanations that accompany it, the answers given to journalists, and how far they depart from expectations formed before the announcement.
Predictability and credibility of communications
For much of the twentieth century, central banks favored a culture of discretion in their communications. In recent decades, however, transparency has become one of the tools of monetary policy. Explaining decisions helps guide expectations, strengthen the institution’s credibility, and reduce the risk of disorderly market reactions.
The effectiveness of communication can be assessed along two dimensions. The first is monetary policy predictability : Given the messages preceding a central bank’s decision, the greater the similarity between those messages and the decision that is ultimately taken, the more predictable monetary policy is. Therefore, it is also more transparent, because the market was better able to anticipate that decision.
The second dimension is central bank credibility : When a central bank makes a decision that surprises the markets (in other words, the similarity between that decision and the messages following it is low), it is necessary to examine the behavior of prices in financial markets. The more prices move in a direction consistent with the central bank’s decision, the more credible that central bank is. If, instead, prices move in a direction at odds with the central bank’s objective, its reputation is relatively weak. A striking example of a monetary policy move that was unexpected yet simultaneously demonstrated the high credibility of the central bank—or of the central banker leading it at the time—is Mario Draghi’s famous “whatever it takes” statement of July 2012.
The literature has already shown that the words of monetary authorities can influence expectations regarding interest rates, exchange rates, stock prices, and macroeconomic conditions. More recently, researchers have begun using social media to observe, almost in real time, the attention, sentiment, and opinions surrounding monetary policy announcements as well.
The study by Masciandaro, Romelli, and Rubera directly compares the content of conversations on Twitter (now X) with that of official announcements, to investigate whether a marked change in similarity between the two texts signals significant surprise or disagreement, and whether this gap is reflected in financial market movements.
Words that influence volatility
The authors analyzed scheduled announcements by the ECB, the Federal Reserve, and the Bank of England between January 2011 and February 2020. The period ends before the pandemic to exclude the extraordinary measures adopted by central banks during the public health emergency. The dataset includes 89 ECB statements and the same number of press conferences, 73 Federal Reserve statements and 41 press conferences, as well as 96 Bank of England decisions. A total of 228,348 English-language tweets identified as relevant were used in the analysis.
Using a natural language processing algorithm, Doc2Vec, the researchers transformed both groups of texts into numerical representations and calculated their degree of similarity. For each event, they then compared two values: the similarity between tweets posted in the five hours before the event and the official announcement, and the similarity between the announcement and messages posted in the 15 minutes afterward. The absolute change between the two values was interpreted as an indicator of the degree of surprise or disagreement generated by the decision.
This measurement was then compared with financial data recorded minute by minute. The analysis covered the main European, US, and British stock market indexes and government bond yields at different maturities.
The results show that, in the case of the ECB, the change in similarity between tweets and the announcement is not associated with greater fluctuations in stock prices around the 1:45 p.m. statement, which summarizes the decision without fully explaining its rationale. But the picture changes during the press conference: The greater the change in similarity, the higher the minute-by-minute volatility recorded across all the indexes considered in the study, from the CAC 40 to the DAX, from the FTSE MIB to the IBEX, as well as the EURO STOXX 50 and the euro-area banking index.
During ECB press conferences, moreover, a greater change in similarity is also associated with a larger overall shift in prices between the periods before and after the event. However, this second relationship does not emerge in US or British stock markets.
Bond markets also react to social media messaging. A greater change in tweet similarity is associated, to varying degrees, with higher volatility in the yields of the benchmark government bonds relevant to each of the three central banks.
The information that matters
Even before serving as a channel through which central banks communicate their decisions, social media is where those decisions are anticipated, interpreted, and challenged. And the gap between what the public expects and what it understands after the announcement leaves a visible trace in the markets.
The difference between ECB statements and press conferences is particularly significant. Volatility does not appear to depend solely on the decision being announced, but also on the additional information provided to explain it. Journalists’ questions and the president’s answers can clarify the central bank’s strategy, but they can also reveal unexpected elements and force market participants to revise their assessments quickly.
For central banks, the results of this research show that online conversations can provide a new indicator of the degree of surprise or disagreement generated by monetary policy decisions, and, more broadly, whether they are communicated effectively. However, it is not possible to determine what causes the observed change.
From the perspective of businesses, investors, and financial market participants, the relationship with volatility and the magnitude of price movements also suggests that these indicators could complement traditional measures of market expectations and reactions.
Donato Masciandaro , Davide Romelli and Gaia Rubera , “ Monetary Policy and Financial Markets: Evidence from Twitter Traffic .” The B.E. Journal of Macroeconomics . 2026; 26(1): 191–249. DOI: https://doi.org/10.1515/bejm-2025-0070 .
