Experts Examine How New Technologies Are Transforming Inflation Forecasting

The TRIM Center hosted a roundtable discussion titled “The Present and Future of Inflation Forecasting.” It brought together participants from the financial sector, academia, and relevant government agencies.

Experts from the TRIM Center presented their research. Nadezhda Yurchenko introduced a BVAR framework. The model’s main strength is its ability to support scenario-based analysis, though its point forecast accuracy may lag behind more straightforward statistical approaches such as optimized ARIMA specifications. Ruslan Chernenko presented an ensemble learning approach designed to maximize predictive performance. While the approach is optimized for accuracy, it offers limited capacity for explicit scenario decomposition.

The presentations framed a broader discussion around several key themes:

Inflation measurement as the foundation of forecasting
As economic structures evolve, so too must the way inflation is measured. The rapid expansion of online marketplaces illustrates this shift clearly, as digital transaction data enables near real-time tracking of price dynamics. Ensuring measurement accuracy in this context requires methodologies that account for ongoing structural changes in the economy—an increasingly important challenge for statistical agencies in Russia.

Forecasting is as much about explanation as prediction
Inflation dynamics are often shaped by discrete supply-side shocks—ranging from the conflict in the Middle East to disease outbreaks and logistical disruptions. While such events cannot be predicted with precision, analysts must explain how such shocks affect inflation across different scenarios.

The forecasting toolkit over the next 3–5 years will go beyond traditional modeling 
The forecasting frontier is expected to center on ensembles that combine econometric models, machine learning systems, and large language models (LLMs) adapted for time-series applications. As the toolkit becomes more sophisticated, the analyst’s role will increasingly shift toward interpretation and communication. In this environment, the ability to clearly explain model outputs to stakeholders will be as critical as forecasting accuracy itself.