Why do people rush out en masse to buy buckwheat, start hoarding foreign currency, or suddenly fear rising prices? The reason often lies not only in the economy itself, but also in stories that spread rapidly through the news, social media, and search queries. Researchers at the Gaidar Institute’s Monetary Policy Department have proposed a new way to track such “viral” economic stories using artificial intelligence and models typically used to study epidemics.
Economists have long known that people do not make decisions like ideal analysts. Most people do not read complex Central Bank reports or study statistics. Instead, people rely on headline news, online discussions, and general sentiment. If everyone around them says that prices will soon rise sharply, many start stocking up in advance—and this in itself can trigger inflation.
Scholars refer to such widespread stories as economic narratives. This term was actively developed by Nobel laureate Robert Shiller. Essentially, a narrative is a simple and emotional explanation of a complex economic situation. For example: “Corporations are to blame for everything,” “The dollar is about to skyrocket,” or “We need to buy urgently before prices go up.” Such ideas can spread almost like a virus
The researchers decided to test whether it is possible to detect the emergence of such “information epidemics” in advance. To do this, they studied Russian Google search queries from 2013 to 2025. The scientists analyzed how people searched for information about inflation, the key interest rate, financial markets, and the Central Bank’s actions.
The most unusual aspect is the research method. To analyze economic news, researchers used the SIR model, which is typically employed by epidemiologists to study the spread of diseases. In this model, some people have not yet encountered the information, others have already been “infected” by the topic and are actively interested in it, while a third group is gradually losing interest.
It turned out that economic news does indeed behave like viruses. For example, reports on Central Bank decisions and changes to the key interest rate spread almost instantly: interest in them skyrockets within a matter of days. Topics like inflation or the state of financial markets, however, “gain momentum” more slowly—people need time to process more complex issues.
This is particularly important for forecasting inflation. If the idea of inevitable price increases begins to dominate public discourse, people change their behavior: they start buying more, stockpiling, and spending money faster. As a result, expectations themselves begin to influence the economy.
“We have seen that the spread of economic narratives can be measured and predicted almost as accurately as the spread of epidemics. This approach could become a new tool for economists and government agencies. Analyzing search queries and the news landscape will help identify alarming public sentiment more quickly, forecast inflation more accurately, and understand which topics most strongly influence the behavior of people and markets,” said one of the study’s authors, Pavel Trunin, Leading Researcher at the Gaidar Institute.
In effect, the researchers propose creating a kind of “public sentiment thermometer” that can indicate in advance when the economy begins to react not so much to actual events as to mass fears and expectations.