This paper describes a new nowcast model for the monthly percent change in the fruit and vegetables price index. Due to the high volatility of this index, the prices of fruit and vegetables have a large effect on the dynamics of the headline Consumer Price Index. Therefore, a standalone prediction for the fruit and vegetables price index may be beneficial. The new model uses multiple data sources, including granular country-wide retail price and wholesale price databases, to precede the official index by two weeks. The main challenge in this prediction task is the small amount of data points available for training and evaluation. Out-of-sample evaluation shows that the new model improves upon the previous one by 33% RMSE using the months May 2020– April 2021 as test periods.

For a PDF of the full article