Building a Real-Time Price Index Based on Prices in Online Stores

Cost and quality of information are becoming the most important characteristics of data for statistical institutes – mastering new methods can deliver larger samples with lower latency at reduced costs. Disadvantages of traditional methods of gathering price data from retailers result in substantial costs in terms of human resources, publication lags, and inaccuracies caused by the human factor. Alternative tracking technologies, on the other hand, require automated product type recognition in order to broaden the sample. This paper suggests a possible solution which addresses the shortcomings of traditional methods. The main task while building a price index based on data from online stores (Real-Time Price Index, or RTPI) was to classify products from the sample by Rosstat’s product categories in order to implement current methods for calculating the consumer price index (CPI). Authors also considered the possibility of facilitating inflation forecasting by lowering the index’ variance with an increased number of tracked products. Authors calculated RTPI based on prices from online stores and then compared it with official CPI. The main hypothesis was not rejected which signifies that there is an inverse connection between the volatility of RTPI and the sample size. Therefore, using online stores’ prices which allows for increased sample size can be beneficial to inflation forecasting.