Reducing biases in the generation of consumer price index with web-based data

Abstract

Governments globally monitor economic well-being using the Consumer Price Index (CPI) to track changes in the cost of living (COLI). However, traditional CPI calculations have issues like substitution bias due to data collection costs. Increasingly available online data and improved computing power are lowering these costs. Nevertheless, transitioning to new CPI methods can create discrepancies in basket composition and weights compared to the traditional CPI. Therefore, maintaining consistency is vital until new methods gain credibility. This paper proposes a new CPI calculation method that aligns with the traditional one. It involves selecting CPI categories, collecting online data more frequently, and determining consistent item weights. The synthetic control method will reconstruct a weighted index of the average prices calculated with the new sources. The expected outcome is a more accurate CPI calculation that reduces errors, adapts faster to new data sources, and improves decision-making. This paper underscores the potential of new web-based data for enhancing CPI calculations in a changing economic landscape.