Investigating the Main Determinants of Inflation in Iran: Application of Hybrid New Keynsian Philips Curve Using Quantile Regression

IntroductionIn investigating effectiveness of monetary policy on the economic stability and control of inflation, the relationship between inflation and real variables is highly important. The Philips curve is one of the most popular relationships in macroeconomic which considers the linkage between inflation and unemployment. It was suggested by Phillips (1958) and expanded by Friedman (1968), Phelps (1968), and Lucas (1973). In the 1990s, New Keynesian Philips Curve was formulated based on nominal rigidity and rational expansions and was widely used in structural models of inflation dynamics and analysis of monetary policy. Despite having a conmonly accepted theoretical background, there have been contradictory results regarding its empirical validity.
High inflation rate has been adversely affected, so it is important to investigate the main determinants of inflation. There are some studies that have investigated dynamics of inflation in Iran. Among them, some have examined the NKPC or hybrid NKPC in Iran. However, these researchers have ignored the utilization of quantile regression method. Quantile regression, proposed by Koenker and Bassett (1978), is an effective tool to overcome the weaknesses in mean regression. In comparison with traditional approaches to estimate HNKPC, quantile regression has two advantages. Firstly, the effects of symmetry and asymmetry of the Phillips curve are studied across quantiles. For example, the effects of explanatory variables may be the different between upper and lower quantiles, which is so important for dynamic monetary policy in different economic circumstances. Secondly, the quantile regression can provide more information than the conditional mean of inflation. Hence, it seems to be able to define the effect of explanatory variables on inflation across quantiles considering the circumstances of frequent inflation in Iran.
MethodologyThe relationship between inflation rate and the economic activity such as the output gap described by a Phillips curve. The most commonly used model Philps curve in macroeconomics is the Hybrid New Keynesian Philips Curve(HNKPC) as developed by Garli ande Gertler(1999) The HNKPC is generally used to investigate the effect of looking forward and looking backward components, which can be expressed as:(1) π_t=γ_f E_t π_(t)γ_b π_(t-1)χGAPϵ_t
Where π_t is the rate of inflation, π_(t) is expected inflation for t at time t, GAP is output gap (i.e., a proxy for marginal cost of production) Whereas Iran is a small country and has open economy, exchange rate also has an impact on inflation rate, so we used the change of the exchange rate on Eq.1. In recent years, quantile regression has been used to estimate Philips Curve (e.g., Boz, 2013; Chorteas, Magonis and Panagiotidis, 2012; Xu, Niu, Jiang, and Huang, 2015).
In recent years, quantile regression has been widely used in many important areas such as economic analysis, financial risk management, and environment modeling (Xu, 2015). Estimation methods of conditional quantile functions were discussed in Koenker and Bassett (1978), where a simple asymmetric version of the sum of absolute errors was minimized. (2) min_(βϵR^k ) [∑_(i∈{i:y_i≥x_i β})▒τ|y_i-x ́_i β_i | ∑_(i∈{i:y_i