Inflation targeting and inflation expectations in Mexico

Abstract

In this article we analyse inflation expectations in Mexico. After a review of the theoretical and empirical literature, we apply unit root, normality and cointegration tests to the data provided by Banco de México (Banxico) in the Survey on the Expectations of the Private Sector Economics Specialists. Our results reject the null hypothesis of normality for inflation expectations over the period 2004:01–2011:12. The exchange rate has become one of the most relevant variables in the transmission mechanism of monetary policy in a small open economy. In this regard, we show the existence of a long-run relationship between nominal exchange rate and interest rate where inflation expectations matter for long-term dynamics.

Keywords:

• inflation expectations
• exchange rate
• Mexico
• inflation targeting

JEL Classification:

• C12
• E31
• E58
• F41

Acknowledgements

This article has been funded by the Vice-Rectorate of Research of the University of the Basque Country into the program ‘Personal Investigador en Formación’, and the Basque Government (Consolidated Research Group GIC10/153). We thank the useful comments of two anonymous referees. The usual disclaimer applies.

Notes

¹ Chiquiar et al. (2010) conclude that since 2001, inflation in Mexico has switched from a nonstationary to a stationary process, with inflation randomly fluctuating around a well-defined mean.

² The first available data for Bolivia is the inflation forecast for the year 2003 published by the WEO database in October 2002.

³ LLC null hypothesis indicates the presence of common unit root. LLC allows to specify intercept, time trend, error variance and to include higher serial correlations for each individual. IPS null hypothesis indicates individual unit root, which allows specifying different AR(1) coefficients for each series. Moreover, IPS captures individual characteristics allowing serial correlation, heterogeneous dynamics and error variances for each individual.

⁴ KPSS null hypothesis is stationarity around a trend. KPSS divides a series into determinist trend, random walk and innovations. Random walk is composed by its value in previous period and an error term, under the null of variance equal to zero in the latter.

⁵ Our results are not entirely conclusive since the validity of the KPSS test is asymptotic and our sample size is small.

⁶ The JB test does not perform well in small samples, leading to reject the hypothesis of a normal distribution in cases where this normal distribution really exists (Deb and Sefton, 1996). However, our result is the opposite, with JB test accepting the existence of normality. Consequently, we can accept the validity of this result.

⁷ Taguchi and Kato (2011) emphasize the difficulties generated by the exchange rate in the implementation and working of IT in emerging markets economies, focusing in East Asian economies. Arestis et al. (2011) also emphasize the role played by the exchange rate in Brazil, another IT country.

⁸ In De Mello and Moccero (2009), inflation expectations in Mexico are integrated of order 1. Unit root presence in the inflation expectations series and the rejection of the null hypothesis of normality could be related to the frequency of the data and the length of the sample. It is possible that with a lower frequency and a longer sample the results change. However, the unavailability of data has made us to use this alternative.

⁹ The only exception is November 2009, where inflation expectations for 1–4 years reached 4.02%.