Построение месячного индикатора и наукастинг номинального ВВП с использованием данных банковской статистики.

Прикладная эконометрика. 2026. Т. 82. С. 105–123.

Финагин М. И., Мирошниченко Д. В.

We construct a monthly indicator and produce a nowcast of nominal GDP in Russia using operational
data. The data used are bank statements and the ruble price of Russian oil. This approach
allows for the nowcast of nominal GDP ahead of official statistics by 3–5 months. Nominal GDP
data is published by Rosstat only on a quarterly basis and with a significant delay after the reporting
quarter. Therefore, high-frequency and operational assessment of the indicator is important for assessing
macroeconomic dynamics. It is also relevant for analysing sectors such as fiscal, financial,
and external. We use statistics of the 0409101 form “Revolving statement of accounting accounts
of a credit institution” as bank reporting data. To assess the quality of the models, we consider indicators
of the indicator’s variability over time and the accuracy of the GDP. Model performance accuracy
is assessed using both actual and vintage data. The paper uses models of temporal disaggregation,
MIDAS, and the DFM model.

Keywords: ВВП, макроэкономическое прогнозирование, наукастинг, Dynamic factor model (DFM), GDP nowcasting, GDP forecasting