Page 28 - Azerbaijan State University of Economics

Vugar Rahimov, Nigar Jafarova: The Exchange Rate Pass-Through to Aggregate
Consumer Price Index and Its Components In Azerbaijan

We include trading partners’ CPI to capture the effects of foreign prices shocks.
According to Purchasing Power Parity Hypothesis, price differences among trading
partners determine the exchange rate in the long run. By including this variable, we
can net out the influence of trading partners’ CPI on the exchange rate.

NEER is included to identify exchange rate shocks. By including both oil revenue
and trading partners’ CPI, we separate their effects on the exchange rate. Thus, the
exchange rate shock can be interpreted as a shock that is isolated from the influence
of those variables.

In our identification scheme the last variable is CPI and its components. It is
obviously included to measure the degree of the exchange rate pass-through to
inflation. Hence, we expect CPI and its components to react positively to NEER
depreciations and vice versa.

In fact, one may try to identify the exchange rate shocks by employing only two
variables (domestic CPI and exchange rate) in the above scheme. However, this
identification scheme violates the ceteris paribus assumption of the impulse
response analysis. As long as we do not include Oil revenue or trading partners’
CPI in the model, there will be only two shocks in the system: exchange rate and
CPI shocks. Such identified shocks will also reflect previous omitted (oil revenue
and trading partners’ CPI) shocks.
This is due to the fact that, for instance, potential effects of oil revenue on CPI (and
its components) do not only work through NEER channel, but also through direct
channel (fiscal channel) (Karimli et al., 2016). If the observed NEER shock is
because of the oil revenue shock, we would expect that CPI shock will also move as
it is contaminated with the oil revenue shock. Therefore, any counterfactual analysis
with the NEER shock will not produce ceteris paribus result. Thus, in our proposed
scheme we include those two variables (oil revenue and trading partners’ CPI) to
avoid the violation of ceteris paribus assumption.

5. RESULTS AND DISCUSSION
In this section we report the empirical results. The desired lag order of the model is
two. The stability tests suggest that all models are stable. The estimates of the
cumulative pass-through coefficients are derived from orthogonalized impulse
response functions. We obtain pass-through coefficients by dividing cumulative
change in price index by the cumulative change in nominal effective exchange rate:

, + = , + / , +1

28