Background:

It is well documented that inflation persistence in Mexico has experienced an unstable behavior through time at the conditional mean distribution. However, its behavior at conditional quantiles of the distribution have been not explored.

Methods:

This study determines the periods in which inflation persistence in Mexico presented structural changes in its conditional distribution using a quantile regression approach. Additionally, the article examines for each period, if inflation follows a stationary behavior using the Quantile Kolmogorov-Smirnov test, estimate the persistence of inflation shocks, and analyze if inflation is converging to the long-term inflation target of 3% impose by the Central Bank.

Results:

The episodes found coincided with periods when Mexico’s economic policies underwent drastic changes that altered the price formation process. The evidence indicates that inflation shocks present an asymmetric behavior, while high magnitude negative shocks rapidly vanish, high magnitude positive shocks tend to be long lasting. Inflation converged to a stationary process in all its conditional quantiles under the inflation targeting regime. Besides, since 2009 the hypothesis that inflation adjusted for seasonal effects remains within the range variability of ± 1% point of the long-term inflation target of three percent cannot be statistically rejected.

Conclusion:

The quantile regression is a useful and convenient statistical tool to analyze inflation persistence. In particular, it gives a clear picture about the periods in which inflation changed, and on the impact of inflation shocks in a specific quantile.