Page 54 - Azerbaijan State University of Economics

period of January 1991-November 2011. The authors first check whether the variables
Granger-cause each other.

Vugar Rahimov: Relationship between PPI and CPI in Azerbaijan: A Wavelet Approach

They later decompose the series to further deepen the analysis through a continuous
wavelet approach. They show the existence of cyclicality and lead-lag relationships
between the variables at different frequencies. In another study, Tiwari et al. (2014)
check the lead-lag relationship between CPI and PPI in Mexico for the period from
January 1981 to March 2009. According to the results, there are bidirectional
relationships between the series as, for example, in longer time scale (8-32 months),
the PPI is leading, while in a short time scale (1-7 months), the CPI is leading.

Another frequency study between producer price index and consumer price index,
using wavelet-based approach, has been conducted by Khan et al. (2018). They use
monthly data over 1999-2016 for the Czech Republic. Additionally, the authors
employ an exchange rate variable as a controlling variable. The findings reveal that
relationships exist between the two series at short-term (higher frequencies).
However, addition of exchange rate increases the time horizon of causality which
shows the sensitivity of the price indices to external shocks in the Czech Republic.

Islam and Kulkayeva (2022) examine the topic in the case of Kazakhstan, which is also a
resource-rich country like Azerbaijan. They use monthly data from January 2011 to
December 2021. To find causality and reveal the relationship time-frequency domain,
Toda-Yamamoto and wavelet approaches have been applied, respectively. While the
causality test indicates a one-way causality from manufacturing producer prices and food
producer prices to consumer prices, the wavelet approach suggests another pattern.
Despite producer prices leading consumer prices in the short term, consumer price is a
leading indicator of producer prices for a relatively longer period.

In a recent paper, Živkov et al. (2023) applies wavelet coherence to investigate the
relationship between consumer prices and producer prices in eight emerging Eastern
European countries, namely, the Czech Republic, Estonia, Hungary, Latvia, Lithuania,
Poland, Slovakia and Slovenia. The sample covers the period from January 1998 to March
2022. They find that there is coherence between the variables in relatively longer horizons
and high coherence is especially apparent during the crisis periods, such as the Global
Financial Crisis and COVID-19 pandemics. Using wavelet-based Bayesian quantile
regression, they also reveal that there are bilateral spillover effects between producer and
consumer prices in all countries excluding Poland and Hungary.

54