By Stefan Gebauer, Thomas McGregor, Julian Schumacher, European Central Bank economists (from the ECB blog)
CENTRAL BANKS They must communicate clearly. This helps the public understand the rationale behind monetary policy decisions and shape market expectations. Their statements sometimes provide clear forward guidance on the future direction of monetary policy.
These statements also reveal how policymakers assess the economy, which in turn influences public expectations about how central banks might react in the future.
Can central bankers’ choice of words – the tone of their communication – affect the economy? In this blog, we show that the tone of their monetary policy communication ECB affects macroeconomic outcomes in the euro area.
Measuring communication tone
OUR ANALYSIS based on a measure of central bank communication. A natural language processing algorithm quantifies the tone of the Governing Council’s policy announcements, in particular the ECB’s monetary policy statements (MPS) and press conference transcripts, which include questions from journalists and answers from the President. These texts are broken down into individual messages on specific topics, to which we assign numerical scores measuring their tone.
THE ALGORITHM is based on a set of dictionaries. First, the individual messages shared during the press conference are categorized by topic, distinguishing between monetary policy, economic outlook and inflation.
These messages are then quantified in terms of the direction and strength of the emotion of the communication.
For example, the sentence “most measurements of the subject inflation “fell further” is rated -1 (dove), while expressions like “underlying inflation has declined rapidly” would yield a rating of -1.5 (even larger margin), with the index ranging from -2 to +2.
THE CLASSIFICATIONS the resulting sentiments are closely associated with “hard” measures of the relevant macroeconomic and financial variables. Changes in the tone of inflation tend to lead to changes in the 1-year ILS forward rates – a common indicator of inflation expectations.
Likewise, changes in the economic tone tend to lead to GDP growth, suggesting that the sentiment contained in the ECB’s official announcement is more than just a reflection of the actual data.
Furthermore, the tone of monetary policy communication is closely aligned with market-based monetary policy expectations, as expressed in the 1-year OIS rate.
The exceptions here are for periods when the central bank applies forward guidance, as was the case for the ECB between July 2013 and mid-2022.
HOWEVER, we cannot simply relate changes in sentiment scores to future economic outcomes: causality can run in both directions. Changes in the political climate are influenced by the monetary policy decisions of the Governing Council, which are influenced primarily by the economic outlook.
THE INDEX Monetary Policy Sentiment Index measures the tone of central bank communication on monetary policy, with an increase in the index signaling a more “aggressive” tone and a decrease in the index signaling a more “soft” tone.
This causality could run both ways: we know that central bank communication generally follows economic data. Therefore, changes in the tone of monetary policy could simply reflect a change in communication caused by changes in economic conditions.
In this case, the updated announcement should not come as a surprise to the public and should have little impact on the financial decisions made by households and businesses.
Identify unexpected changes in communication
THE TONE However, communication can also provide new insights into the direction of monetary policy itself. What happens when changes in monetary policy communication surprise the public? To answer this question, we construct a measure of the extent to which central bank communication contains surprises, that is, changes in the tone of policy that are independent of current and expected macroeconomic conditions.
WE USE two steps, to isolate what we call “monetary policy communication shocks” from the expected tone of monetary policy communication by the public.
THE FIRST The first step is to measure the extent to which changes in the tone of monetary policy are due to changes in the Eurosystem inflation and growth forecasts available at a given Governing Council meeting.
THE SECOND The next step is to limit communication surprises to those in which the financial market reaction shows a pattern consistent with a monetary policy shock. For example, an aggressive change in policy tone would have to coincide with a fall in stock prices to qualify as a political communication shock. A change in communication would have to occur in conjunction with a rise in stock prices to qualify as a political communication shock.
RESULTS show that surprises in central bank communication significantly affect prices and real activity. A sudden and aggressive change in monetary policy communication (positive shock) leads to a significant and noticeable reduction in inflation and economic activity.
WE CONFIRM The typical result is that interest rate shocks have large effects on the economy. In addition, we find that communication shocks are also important. Their effect is smaller, but still significant. Furthermore, the main impact on the economy is achieved later by communication shocks.
One reason for the lagged effects of communication shocks could be that observers need more time to internalize changes in central bank communication on monetary policy and factor them into economic decision-making.