Innovative Forecasting of U.S. Inflation: A Deep Dive

A new NBER study shows that multivariate forecasting models of inflation have proven more accurate than univariate approaches. This highlights the necessity for a detailed analysis of different economic sectors to better comprehend price trends in the post-pandemic environment.

The impact of the COVID-19 pandemic on global economies, particularly inflation trends, has been significant. In 2021, a persistent rise in prices in the United States caught economic institutions and policymakers off guard. Initially viewed as a transient phenomenon, this surge led to substantial forecast errors throughout late 2021 and early 2022. A new study from the NBER investigates whether univariate time series models, which rely solely on historical inflation data, were capable of accurately predicting such trends. The findings indicate that multivariate approaches, incorporating sector-specific inflation measures, performed better. This method suggests a more robust way to understand inflation behavior in the post-pandemic era, especially given the volatility and varying persistence observed.

Source: NBER Working Papers