The measure matters: differences in the passthrough of inflation expectations in Colombia

This paper was produced as part of the BIS Consultative Council for the Americas (CCA) research conference on "Growth, productivity and macro modelling in the Americas", held in Ottawa on 26–27 October 2023.

Summary

Focus

Our study examines how different measures of inflation expectations influenced inflation dynamics in Colombia from 2009 to 2024. We estimate New Keynesian Phillips curves (NKPC) and structural VAR (SVAR) models using data from economic surveys and sovereign bond yields. We explore differences in how these expectations are formed to explain our findings.

Contribution

Monetary authorities increasingly rely on various measures of inflation expectations for policy analysis. Understanding the different effects of these measures helps central banks implement policies that avoid unintended consequences, such as unnecessary contractions in economic activity.

Findings

Our findings indicate that survey-based expectations have a greater impact on inflation than financial market-based expectations. A 1 percentage point increase in survey-based expectations leads to a 0.8 percentage point rise in inflation, while the same increase in market-based expectations results in a 0.67 percentage point rise. These differences are linked to how different economic agents form their expectations, influenced by asymmetric losses, forecasting costs and information rigidities.

Abstract

This study examines the effect of different measures of inflation expectations on inflation dynamics in Colombia from 2009 to 2024. We estimate New Keynesian Phillips Curves (NKPC) and Structural VAR (SVAR) models using data from economic surveys and sovereign bond yields. Our results show that survey-based expectations have a greater passthrough to inflation, with a one percentage-point increase leading to a 0.8 percentage-point rise in inflation, compared to a 0.67 percentage-point rise from market-based expectations. These differences are attributed to how economic agents form expectations, influenced by asymmetric losses, forecasting costs, and information rigidities. Our findings provide crucial insights for monetary authorities, who increasingly rely on various measures of inflation expectations for policy analysis. Understanding the distinct effects of these measures helps central banks implement policies that avoid unintended consequences, such as unnecessary contractions in economic activity.

JEL Classification: C26, D84, E12, E31

Keywords: inflation expectations, inflation dynamics, new-Keynesian Phillips curve, generalised method of moments