Inflation Forecasting with Inflation Sentiment Indicators
Roland Döhrn, Christoph Μ. Schmidt, Tobias Zimmermann
Abstract
Open-access reader
Roland Döhrn, Christoph Μ. Schmidt, Tobias Zimmermann
Abstract
Open-access reader
In this paper we argue that future inflation in an economy depends on the way people perceive current inflation, their inflation sentiment.We construct some simple measures of inflation sentiment which capture whether price acceleration is shared by many components of the CPI basket. In a comparative analysis of the forecasting power of the different inflation indicators for the US and Germany, we demonstrate that our inflation sentiment indicators improve forecast accuracy in comparison to a standard Phillips curve approach. Because the forecast performance is particularly good for longer horizons, we also compare our indicators to traditional measures of core inflation.Here, the sentiment indicators outperform the weighted median and show a similar forecasting power as a trimmed mean. Thus, they offer a convincing alternative to traditional core inflation measures.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper we argue that future inflation in an economy depends on the way people perceive current inflation, their inflation sentiment.We construct some simple measures of inflation sentiment which capture whether price acceleration is shared by many components of the CPI basket. In a comparative analysis of the forecasting power of the different inflation indicators for the US and Germany, we demonstrate that our inflation sentiment indicators improve forecast accuracy in comparison to a standard Phillips curve approach. Because the forecast performance is particularly good for longer horizons, we also compare our indicators to traditional measures of core inflation.Here, the sentiment indicators outperform the weighted median and show a similar forecasting power as a trimmed mean. Thus, they offer a convincing alternative to traditional core inflation measures.
Key concepts: Inflation (cosmology), Economics, Core inflation, Econometrics, Construct (python library), Phillips curve, Predictive power, Macroeconomics