Asymmetric inflation persistence in Latin America: evidence from Chile, Colombia and Peru using quantile autoregression analysis

Abstract

This paper investigates asymmetric inflation persistence in Chile, Colombia and Peru using quantile autoregression on monthly data from 1992–2023. Unlike conventional approaches that assume uniform adjustment speeds, this method captures heterogeneous dynamics across the inflation distribution. Results indicate global stationarity but marked asymmetries: positive shocks display substantially greater persistence than negative ones. The unit root hypothesis cannot be rejected at and above the 60th, 70th, and 80th quantiles for Colombia, Chile, and Peru, respectively, implying that high-inflation episodes endure while negative deviations dissipate quickly. Robustness checks controlling for multiple structural breaks show that persistence declined significantly following the adoption of inflation-targeting regimes and improved macroeconomic management, yet asymmetric patterns remain. Findings are robust to alternative steady-state specifications. The evidence supports asymmetric monetary policy responses and offers practical guidance for central banks. Extended analysis for Brazil confirms the prevalence of these asymmetric dynamics.