 |
| Twitter tracking consumer sentiment on pricing |
The Bank of Italy said on Monday that a number of empirical indicators, based on the content of millions of tweets on its Twitter platform, closely track consumer price perceptions and leave room for powerful new monetary policy tools.
Economists around the world are increasingly turning to social platforms and other unconventional sources to assess consumer behavior efforts amidst persistent inflation.
The researchers found that these indicators are in line with final inflation values and the current measurement standards for price expectations of the Italian National Statistics Office and financial markets, and are in real time and provide more accurate detail.
The author of the 107-page research report said: Study results indicate that Twitter may be a new resource that can find a way to inculcate faith, and we believe Italy-focused research can be replicated elsewhere.
Twitter has nearly 200 million monthly active users worldwide and there were about 10 million active users in Italy in 2019.
The analysis collected for the first time 11.1 million tweets published in Italian between June 2013 and December 2019 that contained at least one set of pre-selected words about inflation, price and price.
The study states that: The number of tweets below comes from an intuitive idea: the more topics people talk about, the more likely they will reflect their views, and their views will influence the expectations of others.
By measuring the daily volume of Tweets that contain predefined keyword combinations, two indicators are generated from this dataset to predict an increase or decrease in inflation.
The most recent indicators are created based on the difference between these two indicators. The author said: This work confirms the importance of receiving information through social media and its political importance.
They also noted that due to the rapid spread of events on social platforms, the metrics adopted on Twitter differ for some cases, recognizing the need for more research to interpret the data.