It can be predicted inflation using Twitter? According to the Bank of Italy, yes. A study curated by four Via Nazionale analysts (Cristina Angelico, Juri Marcucci, Marcello Miccoli and Filippo Quarta) compared three indicators of inflation expectations by consumers: the first type, daily, is “constructed with techniques of machine learning algorithm and text analysis based on social network messages Twitter”; the second, monthly, is based on surveys by theIstat; the third, again daily, is linked to the prices of derivatives on inflation (inflation swaps). The last category of indicators, the authors point out, measures expectations "only indirectly", because it includes "an unobservable risk premium".
The analysis was carried out using "the most recent specific algorithms for the treatment of large volumes of data (Big Data Analytics)” and demonstrates that Twitter-based indicators “have a significant correlation with other existing measures of inflation expectations”. Not only that: projected on a monthly basis, the data provided by the social network “have a predictive power” even “superior to that of the other information sources available” and “therefore represent a complementary measure available in real time”.
If the data is analyzed with correct scientific methods, then, Twitter is a trusted source from which to obtain indications on the future trend of prices. The affirmation is not idle, because – when it comes to inflation – the more sources you have available, the better. In fact, Istat calculations are available only on a monthly or quarterly basis, while those based on derivatives return market expectations more than those of consumers. The analyzes obtained from Twitter, on the other hand, have the merit of combining timeliness and accuracy.
After all, as the Bank of Italy analysts underline, the 140-character social network can count on “200 million active users per month all over the world and about 10 million active users in Italy in 2019”. As a result, discussions on this platform “bring to light the opinions of the community on various issues of politics, technology, economics and so on” and represent “a unique opportunity for researchers interested in studying consumer beliefs.
