ABSTRACT
Despite extensive research and interest in the causality between inflation and inflation uncertainty, no firm conclusions have been drawn. This study uses an improved meta-regression analysis, namely a meta-Granger causality testing, to address two main questions: (Q1) What are the factors that explain the large differences in the results? We show that much of the variation between studies can be due to overfitting bias in addition to the variety of model specifications, different data sources, diverse inflation uncertainty indicators, and dissimilar country-specific characteristics. (Q2) Is there evidence of genuine effect between inflation and inflation uncertainty? We find that there is a robust genuine effect from inflation uncertainty to inflation. We cannot provide evidence of a genuine causal effect from inflation to inflation uncertainty, although excess significance is present. This excess significance is dominantly explained by overfitting bias. We also deduce that the results may suffer from omitted-variable biases obscuring the true causal effects.
JEL Classification: B22 ; E31 ; C10 ; C83.
