Estimating inflation persistence by quantile autoregression with quantile-specific unit roots

Testing for asymmetric adjustment in weekly Brazilian inflation

“…Our main fi ndings indicate that we cannot reject the unit root at the 5% level in a broader set of upper-tail quantiles (τ = [0.75;0.95]) relative to Gaglianone et al (2018) and Maia and Cribari-Neto (2006). However, there is strong evidence that the weekly Brazilian infl ation rate is not a unit root process at the average and lower quantiles.…”

“…Since our data are free of the pegged exchange rate effects, we speculate that the current monetary policy regime is associated with an exacerbation of asymmetric infl ation dynamics. Further, as the power of the ADF type tests depends upon the sampling frequency and the sample size, we get more precise estimators (lower uncertainty) and higher power of the test relative to Gaglianone et al (2018) and Maia and Cribari-Neto (2006).…”

“…In a study based on Monte Carlo simulation, Choi and Chung (1995) show that, in the case of ADF-type tests, using high-frequency data (weekly observations instead of monthly, for example) can provide a considerable improvement in the fi nite sample power of the unit root test. Hence, we hope to reach more robust and reliable results when compared with studies for infl ation persistence based only on monthly data, as done by Maia and Cribari-Neto (2006) and Gaglianone et al (2018). Bogdanski et al (2001) and Fraga et al (2004) argue that Brazilian infl ation persistence may depend on a series of state-regulated prices like indexed wage contracts, public transport fare, gasoline, Diesel oil, electricity, telephone and post offi ce rates, education, health plan rates, and many other prices (see Figueiredo and Ferreira, 2002, p. 6).…”

“…Maia and Cribari-Neto (2006) applied the QAR model using monthly infl ation data to the period ranging from August 1994 to April 2004. More recently, Gaglianone et al (2018) applied the QAR model to analyze the Brazilian monthly infl ation based on samples ranging from January 1995 to May 2017. Both studies use only the restricted version of the QAR model.…”

“…Because the government might choose to alter almost continuously the exchange rate in a pegged exchange rate regime as means of dampening infl ation in the short run (Dorbunsch et al, 1990), the period under analysis in this paper is focused only on a fl oating exchange rate system. Maia and Cribari-Neto (2006) and Gaglianone et al (2018) use monthly observations that include the pegged exchange rate regime and the fl oating exchange rate system. Our estimates avoid such mixed effects on the infl ation level and its persistence.…”