‘Any Real-Economy Company Can Use Our Products’

© iStock

The HSE Centre for Financial Research and Data Analytics combines fundamental and applied work, including in areas unique to Russia such as the connection between sentiment in the media and social networks and financial markets. The HSE News Service spoke with the centre’s director, Professor Tamara Teplova, about its work. 

— When and on whose initiative was the centre established?

— The centre started as a small research and educational laboratory in 2012. We worked on projects with investment companies and launched a club to discuss investment ideas. Under the guidance of our teachers, students created portfolios and presented investment ideas to market professionals for scrutiny—mainly concerning the Russian stock market.

Tamara Teplova

We soon realised that beyond verification, the ideas required a sound foundation for studying anomalies. We were ambitious enough to pursue basic science and publish articles, so we applied to establish an international laboratory. Following a competition, we received funding under HSE University’s Centres of Excellence project. We are implementing three-year projects that can be extended based on the results of an inspection and evaluation of the commitments fulfilled.

In 2022, the laboratory was reorganised into the Centre for Financial Research and Data Analytics.

— What are the centre’s key activities?

— Investment strategies are based on identifying anomalies. Some are short-lived; others persist for a long time. We therefore study where investors encounter short-term phenomena and long-term trends, when and why abnormal returns may cease, and how to use them to make money.

In 2022, we applied for an RSF grant, which was supported because the topic of financial market sentiment—emotionally charged assessments in the media and social networks—had not been thoroughly studied in Russia: unlike media headlines, social media discussions had largely remained outside the focus of researchers.

We contacted HSE graduates working in China, which gave us access a pool of data and allowed us to participate in analysing large datasets from original social network channels and chats. We learned how to parse messages and various reactions to them, translate them into quantitative sentiment assessments, and use this to built a sentiment metric suite. We have, for instance, obtained patents for investment-related social media jargon and sentiment metrics.

Unique sentiment detection models—relying on the analysis of texts in Russian and Chinese social networks and accounting for jargon, emojis, and other expressions of emotions—enabled us to publish articles in first- and second-quartiles journals of international rankings.

Early-career researchers at the centre and students working on their theses study market sentiment broadly and for specific assets (including cryptocurrencies), build metrics based on daily and intraday data, test their impact on asset behaviour, and try to develop predictive models.

To obtain her Master’s degree in Financial Markets and Financial Institutions, Olga Zelenova defended a thesis on the influence of sentiment on the French market. After studying investment-related social networks in both French and English, she found out that English-language discussions affected stock behaviour, while French-language discussions had no significant influence.

Our work on the Chinese market helped attract external funding, establishing the reputation of our centre among domestic customers. From an academic standpoint, we are particularly interested in the Russian market, for which we have proposed original sentiment metrics. Maxim Fayzulin studied how investors on investment platforms perceived the same stock market events and how sentiment metrics performed. Building opinion-divergence metrics—across different investment platforms on social networks, accounting for both positive and negative feedback—makes it possible to draft one’s own strategy. He is scheduled to defend his Candidate of Sciences dissertation on May 12, 2026.

We’ve come to realise that attitudes on Russian and Chinese social networks and chats differ greatly.

— Can you elaborate?

— As a rule, Chinese users tend to encourage one another. When someone complains about making a bad deal and losing money, others mostly respond with reassurance about future successes: act differently, and you will profit on other deals.

Russian social networks are dominated by mockery: remarks along the lines of ‘why would you leap without looking?’ and ironic commentary about the person who made the mistake. When someone reports a good deal, Russians are more likely to bring them down a peg: there is nothing to brag about—one deal is not a business.

— What other research areas do you pursue?

— One obviously interesting area has partly grown out of our work on sentiment: the crypto market. It is no secret that young people actively work in companies related to digital assets, use cryptocurrencies, and are curious about the factors that can explain value dynamics of cryptocurrencies and NFTs. An open question is how to hedge risks amid today’s high uncertainty. We have concluded that mood and expectations—the perception of current events on social networks—are among the significant factors.

Valeria Baklanova, now a teacher at our centre, defended her dissertation on modelling the Bitcoin and NFT markets in terms of the influence of prominent figures (Donald Trump proved to be the most influential among them). By analysing their statements and reactions to them from a wide range of social network participants, it is possible to understand the dynamics of different crypto market segments.

Our book Crypto Code outlines the principles of cryptocurrencies regulation in different countries—they are prohibited in China, for example, while their use is encouraged in South Korea and Japan—and describes how different investor groups have approached them, and how the emergence of new crypto assets and the arrival of institutional investors have shaped the market.

Another area concerns artificial intelligence—machine learning and neural networks. When we analysed texts, we needed AI tools and algorithms, and we realised it was possible to leverage them to solve a broader set of tasks. We moved towards predictive models, deeper understanding, and crisis prediction. We published an article in a first-quartile journal, which was actively cited by domestic economic and financial editions.

It is also possible to build predictive models and architectures for developing capital markets, which have many more influencing factors than developed ones. This area is the future.

— Which areas would you call the most promising?

— Predictive models, the search for bifurcation points, the identification of new anomalies, the testing of established views, and the identification of the causes of market inefficiency arising from growing conflicts—especially with the use of AI models—are all of great interest. We see many opportunities for engagement here.

We also conduct basic research focused on studying the fundamentals of how markets function and creating tools for monitoring them, which is equally important.

— The centre regularly monitors financial markets. What key indicators do you and your colleagues analyse?

— We closely follow the analytical work of the Bank of Russia and the Moscow Exchange; we appreciate their approach. They have the opportunity to build compelling products using the first layer of superdata. Unlike them, we do not have access to things as intraday exchange indicators or individual participants’ transactions. We work with public data and identify deeper processes, such as a wide range of liquidity indicators on the stock market. No one else calculates these. Sergey Gurov is currently building models for analysing liquidity on the cryptocurrency market as part of an RSF grant.

© iStock

— Who is particularly interested in your developments: scientists, government, financiers, or private investors?

— Probably professional, institutional, and private investors, as well as those interested in the models themselves. We have become proficient at developing them, and they translate well beyond financial markets—any real-economy company can use our products.

One applied study, for example, was commissioned by the Roscosmos state corporation, which uses our models to analyse satellite imagery.

— What role does the centre play at the Faculty of Economic Sciences (FES)?

— We work in the field of financial economics, providing scientific support and data for Master’s programmes in Financial Markets and Institutions, Financial Engineering, and Investments in Financial Markets. Our staff reviews theses, conducts workshops, and offers free summer schools. At the next one, participants will learn how to develop advanced models using Python. We will run several sessions to help prepare graduation papers and promote the development of professional competencies.

For years, we have systematically supported the Higher League Olympiad by creating tests and tasks. This has become quite a challenge: it now takes place online, and some participants attempt to use AI in their solutions. It is a competition between publicly available AI tools and our own. We formulate tasks in such a way that mechanical, unreflective use of AI produces wrong answers, encouraging participants to think for themselves rather than rely on AI-generated solutions. Our staff also takes part in the FES winter school, which features team and individual competitions alongside specialist workshops.

— How are research results incorporated into the educational process?

— We teach several courses: fundamental and technical analysis, cryptocurrency market modelling, a Python course, and AI in financial economics. We supervise students preparing their bachelor’s and master’s theses. We also hold mentors’ seminars, where students can immerse themselves in their chosen research areas and receive informed recommendations and feedback.

— How actively do students and doctoral students participate in the laboratory’s work?

— Very actively. We have only five experienced researchers; the rest are early-career specialists. Some of our graduates combine work with doctoral studies, and some teachers work in banks. Students and doctoral students are in high demand in the financial market, in banks, and at investment companies.

— What research centres do you cooperate with?

— We have close contacts with the departments of the Financial University working in our field. We also cooperate with the RANEPA Laboratory for the Analysis of Institutions and Financial Markets, headed by Alexander Abramov. They build factor strategies on the Russian stock market, and one of them may incorporate our sentiment metrics. Our RANEPA colleagues are interested in our metrics and are working to assess the impact of investor attitudes towards individual companies on systematic risks. 

— How do you manage to maintain international academic cooperation at present?

— After 2022, institutional cooperation has been difficult. We continue to interact with foreign colleagues on a personal basis—which is also why the international laboratory became the Centre for Financial Research and Data Analytics.

Every year we hold an international conference with colleagues from the University of Bahrain, where doctoral students from HSE University and the Middle East present their work.

— Stock trading and the crypto market are often associated with the chance to improve one’s financial situation quickly. What do you tell your students about this?

— The way I see it, a wealthy person is a free person: someone with access to current publications, databases, and a wide circle of like-minded peers. They can express themselves, choose work they find meaningful, think critically, and remain financially independent of their superiors. I tell students that it is entirely possible to pursue research and work in the markets simultaneously. As for financial well-being: I think one should take their time getting rich, rather than rushing.