TITLE:
Random Coefficient Modelling of the Global Effect of Exchange and Monetary Policy Rates on Inflation—A Frequentist and Bayesian Generalized Additive Mixed Model Approach
AUTHORS:
Chioma Okoronkwo, Geoffery Uzodinma Ugwuanyim, Hycinth Chukwudi Iwu, Harrison Obiora Amuji
KEYWORDS:
Inflation, Exchange Rate, Monetary Policy Rate, Frequentist Generalized Additive Mixed-Model, Bayesian Generalized Additive Mixed-Model
JOURNAL NAME:
Open Journal of Statistics,
Vol.15 No.5,
September
30,
2025
ABSTRACT: This research evaluates the effect of monetary policy rate and exchange rate on inflation across continents using both Frequentist and Bayesian Generalized Additive Mixed Models (GAMMs). Extending an earlier study that employed Frequentist and Bayesian Linear Mixed Model, continent-specific random slopes for exchange rate were incorporated to assess the variability in the rate at which changes in exchange rate influence inflation. Fixed effects captured the overall impact of the predictors, while random effects accounted for regional differences. Results consistently showed that the monetary policy rate significantly affects inflation, whereas the exchange rate does not. Strong evidence supported variation in baseline inflation across continents (random intercepts), but findings on random slope variability were mixed, suggesting modest and model-dependent heterogeneity. Bayesian models offered a slightly better fit and predictive accuracy. These findings underscore the central role of monetary policy in inflation control, while exchange rate effects remain context-dependent. These results highlight the importance of accounting for regional heterogeneity when modelling global inflation dynamics. Policymakers should tailor inflation strategies to regional contexts and prioritize robust monetary policy tools over exchange rate management. These findings are associational, not causal and future research should adopt a credible causal identification strategy to establish causal relationships.