The objective of this paper is to examine the general relationship between the year-on-year inflation-rate and nominal interest rates, contributing to a better understanding of what high inflation implies for nominal rates. Although a heavily researched topic, economists still seem far from a consensus regarding the relation between these two variables; some papers suggest a strong relationship – while others reject the idea of an empirically observable relationship altogether. This scientific unambiguity risks leaving policymakers and households reacting with passivity to inflation, although perhaps they should not.
Joining Westerlund (2008) in his criticism, this paper argues that the ambiguity of earlier research is mainly due to the overwhelming focus on single-country or country-by-country analysis. These methods are prone to end up in ambiguity, as some countries or time-periods indicate a correlation, whilst others do not. To avoid such ambiguous results, this relationship needs to be analyzed using panel-data methods. Instead of using the panel-cointegration tests suggested by Westerlund however, this paper argues that many earlier studies – including Westerlund – make a mistake when diagnosing inflation and/or nominal rates as non-stationary variables. This is because such conclusions are at odds with the de-facto mean reverting property of these variables. Noticing these shortcomings in earlier research, this study discards cointegration tests in favor of pooled OLS, fixed effects- and random effects models, fitted to a panel of 31 countries spanning the period 1965-2022. The results suggest close to a 1:1 relationship between the year-on-year inflation rate and the 3-month T-bill rate. When inflation has been high, nominal rates have generally been high as well, and vice-versa.