Visible prices and their influence on inflation expectations of Russian households

A multitude of recent research shows that the inflation expectations of households are far from rational. In making inflation forecasts, people tend to focus on the prices of particular goods and services, which they can observe every day – ‘visible prices’. In this paper, we propose a new method for the identification of such items. Our novel ‘brute force’ algorithm automatically sorts through the full array of prices of goods and services given by Rosstat and constructs consumer baskets. It then selects the best baskets based on their ability to forecast the inflation expectations of Russian households from the FOM Survey. In the end, we get a decomposition of various met-rics of inflation expectations for visible prices which also demonstrates good forecasting perfor-mance (as compared to the AR(1) process as a benchmark). To ensure robustness, we use an alter-native method (optimisation with regularisation) and a variety of metrics of inflation expectations. As a result, we get lists of ‘robust visible items’ which include not only foodstuffs but mainly durable goods and services. Surprisingly enough, oil and petrol, which are typically labelled ‘vis-ible goods’ in research, do not fall into this category for Russia.