Central Bank of Azerbaijan Develops Index for Early Forecasting of Food Inflation

The dynamics of prices for food products imported into Azerbaijan can signal the formation of food inflation in the country approximately two months in advance.

This conclusion is reflected in a study titled "Construction and Forecasting of an Imported Food Price Index for Azerbaijan," prepared by Central Bank of Azerbaijan (CBA) staff members Tamkin Nuriyev, Aygün Garayeva, and Gulzar Tahirova.

As part of the study, an Imported Food Price Index (IFPI) was developed to more accurately track price changes for food products imported into Azerbaijan. The index is based on data from approximately 800,000 customs operations over the period from January 2018 to February 2026. These data were selected from about 16 million raw import transactions.

The new index is designed with the structure of the consumer basket in mind and includes 34 main product categories and 44 product types with a high degree of import dependence. According to the authors, this approach helps reduce the aggregation problem that can arise in traditional import price indicators calculated across broad product groups, and more accurately measures the impact of external price changes on domestic food prices.

One of the main results of the study is the leading nature of the relationship between the IFPI and the official Consumer Price Index (CPI) for food. According to calculations, the correlation coefficient between the two indicators reaches 0.81 with a two-month lag. This shows that changes in imported food prices can signal in advance the subsequent dynamics of food prices in the country.

"Another important advantage of this indicator is the timing of data availability. Customs data become available on average 20–25 days after the reporting month. Data from the State Statistics Committee on food CPI are usually published 10–15 days after the end of the month. As a result, after the IFPI for a given month is formed, this indicator can provide an information advantage of approximately 6–7 weeks compared to the publication of the official food CPI for the two subsequent months," the study notes.

The study also develops an approach to forecasting imported food prices based on machine learning models. The authors, comparing this approach with traditional linear models, report higher forecasting results.

According to the study, changes in global food, energy, and fertilizer prices also significantly influence the dynamics of imported food prices. The correlation between the IFPI and the FAO global food price index was 0.70, with fertilizer prices 0.66, and with natural gas prices 0.62.

The authors note that in Azerbaijan, food products account for about 43 percent of the consumer basket. In this regard, more accurate and timely tracking of external shocks in import prices is important for forecasting food inflation as well as overall inflationary processes.

The study notes that the developed index and its forecasts can be integrated into the CBA's more detailed food inflation forecasting models. It is reported that this approach can contribute to earlier identification of food price dynamics and improve the accuracy of overall inflation forecasts.

According to the authors, in economies with high import dependence, along with monitoring international price indicators, it is important to develop more detailed indicators that reflect how external price shocks are transmitted to domestic prices through specific goods and the consumer basket.