What is the effect of imported inflation and central bank credibility on the poor and rich?

ABSTRACT

This study analyzes two phenomena regarding the inflation of the poor and rich. Firstly, considering that imported inflation is a source of inflation, we investigate whether the harmful effect of imported inflation is higher for the poor than the rich. Secondly, assuming that central bank credibility helps keep inflation under control and that the poor are more susceptible to the damaging effect of the inflation tax, we evaluate whether the credibility is more beneficial to the poor. Based on data from the Brazilian economy, the findings indicate a higher effect of imported inflation on food inflation and poor households. In contrast, central bank credibility effectively reduces inflation for the poor.

KEYWORDS:

• Imported inflation
• inflation of poor and rich
• headline and food inflation
• central bank credibility

JEL CLASSIFICATION:

• E31
• E64
• E58

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Throughout this paper, we consider ‘imported inflation’ as synonymous with the expressions ‘import price inflation’ and ‘external inflation’ usually found in the literature.

² On the impact of food inflation on low-income households, see Ginn and Pourroy (2020) and Fujii (2013).

³ Hottman and Monarch (2020), using data regarding US households from 1998 to 2014, found evidence that imported inflation was higher for lowest-income households.

⁴ For an analysis concerning imported inflation, see Albuquerque and Baumann (2017); Szafranek (2017); Watson (2014); and Gruen, Pagan, and Thompson (1999).

⁵ Table A.1 (appendix) shows the descriptive statistics and sources for all variables in the models.

⁶ Regarding the advantages of using this index compared to others, see de Mendonça (2018).

⁷ Regarding how the market expectations system works, see: https://www.bcb.gov.br/en/monetarypolicy/marketexpectationsfaq.

⁸ de Mendonça and Garcia (2023) provide a similar Phillips curve regarding the analysis of supply shock (oil price pass-through) and central bank credibility.

⁹ In order to simplify our analysis, we are not working with a micro-founded Phillips curve. For a comprehensive review, see Ascari and Sbordone (2014). Specifically for the case of central bank credibility and the Phillips curve, seede Mendonça, Filho, and Souza (2023).

¹⁰ Regarding applications of GMM estimation, see Wooldridge (2001).

¹¹ The list of instruments for each model is available in Table A.3 – appendix.

¹² The COVID-19 pandemic reallocated consumer spending among categories during 2020 (Cavallo 2020; Kantur and Özcan 2021; Lane 2021). According to CBB (2020a), the COVID-19 pandemic has driven inflationary pressure and changes in relative prices in Brazil since March 2020. Hence, we made a robustness analysis of the results in Table 1, providing new regressions considering the pre-pandemic period from June 2007 to December 2020 (see table A.5 - appendix). In general, the findings did not change considerably from those in Table 1.

¹³ For an analysis of the harmful effects of food inflation on the poor, see, e.g. Fujii (2013), Alem and Söderbom (2012), and Cudjoe, Breisinger, and Diao (2010).

¹⁴ It is important to highlight that we use headline inflation for the poor and rich in Table 2. Moreover, the results align with those in Table 1 for headline inflation.

¹⁵ Ipea provides inflation of households that belong to six income levels: very low-income, low-income, low-middle income, middle income, high-middle income, and high-income. To check whether imported inflation and credibility have heterogeneous effects on income groups, we also provided the regression results for the average inflation of the entire group of households – INFM (see table A.6 - appendix). In general, the coefficients on INFM and CRED_k are in line with the expectation of an intermediate effect between the case of poor and rich, thus reflecting a heterogenous effect among income groups.

¹⁶ DINF (the difference between poor and rich inflation) is the result of the difference between INFP and INFR.

¹⁷ Based on the usual lags selection criteria, the models were estimated with two lags. This result is reasonable because, although we are using monthly frequency, inflation is accumulated in twelve months. Furthermore, the VAR stability condition is satisfied (see table A.4 - appendix).