A short term credibility index for central banks under inflation targeting: an application to Brazil

Extracted main text — title through conclusion, appendix excluded. This is what our citation measures are computed over, published so the extraction can be checked by eye.

3,629 characters · 2 sections · 0 citation commands

Rendered from LaTeX for readability, not typeset faithfully. Citation keys are highlighted; maths is left as source; figures, tables and equation environments are summarised rather than reproduced; unrecognised commands are greyed out so nothing is silently dropped. Email addresses are removed.

titlepage\begin{center} { A short term credibility index for central banks under inflation targeting: an application to Brazil } { Alain Hecq\footnote{Maastricht University School of Business and Economics}, João Victor Issler\footnote{FGV EPGE Brazilian School of Economics and Finance} and Elisa Voisin\footnotemark[1]\footnote{Corresponding author : Elisa Voisin, Maastricht University, Department of Quantitative Economics, School of Business and Economics, P.O.box 616, 6200 MD, Maastricht, The Netherlands. Email: [email removed].} } \\ { May 2022 } \end{center} \begin{abstract} This paper uses predictive densities obtained via mixed causal-noncausal autoregressive models to evaluate the statistical sustainability of Brazilian inflation targeting system with the tolerance bounds. The probabilities give an indication of the short-term credibility of the targeting system without requiring modelling people's beliefs. We employ receiver operating characteristic curves to determine the optimal probability threshold from which the bank is predicted to be credible. We also investigate the added value of including experts predictions of key macroeconomic variables. \end{abstract} Keywords: inflation rate, noncausal models, forecasting, predictive densities, probabilities, credibility.\\ JEL. C22, C53

\graphicspath{ {Pictures/} }

\import{./}{1-Introduction.tex} \import{./}{2-MAR_models.tex} \import{./}{3-Predictions.tex} \import{./}{4-Application.tex} \import{./}{5-MARX.tex} \import{./}{6-Credibility.tex}

Conclusion

This paper investigates the probabilities that the Brazilian inflation rate remains within the target bounds determined by the central bank. We estimate mixed causal-noncausal models with leads and lags of the inflation rate and the first noticeable results are the stability of the parameters estimation over time. Contrary to usual practice when employing mixed causal noncausal models, which consists in forecasting bubble bursts, we choose the stable period between 2017 and 2020 for the forecasting exercise. This allows us to forecast at farther horizons using a sampling importance resampling algorithm. We perform 1-, 3- and 6-months ahead probability forecasts. For short-horizons forecasts we correctly track probabilities to stay in the target bounds and interpret those results as the short-term credibility of the inflation targeting system based on historical inflation without the need to model people's beliefs. In longer horizon, probabilities seem to converge to a 50/50 result. We augment the univariate model with key economic variables and results show that during stable times, the addition of those variables do not impact significantly the predictions obtained from the univariate model. We then compare our measure of short-term credibility of the central bank with existing indices in the literature. We find that the distinct horizon and perspective used to build the indices makes our short-term credibility index complementary to the existing long-term indices. We employ receiver operating characteristic curves to determine the adequate probability threshold from which the Central bank is predicted to be credible or not.

\import{./}{8-Appendix.tex}

Acknowledgments

The authors would like to thank Joann Jasiak and the participants of CFE 2021 for valuable suggestions.