Forecasting the CPI of the Kurdistan Region of Iraq Using the Combined ARIMA and GARCH Model With Wavelet Analysis

A wavelet-based hybrid time series model is developed to enhance inflation forecasting accuracy. This study develops a wavelet-based hybrid time series framework to enhance the accuracy of inflation forecasting. The approach integrates Coiflets-based discrete wavelet transform with ARIMA and GARCH models to improve the prediction of the consumer price index (CPI) in the Kurdistan Region of Iraq. The CPI series, comprising 180 monthly observations from May 2008 to December 2023, was decomposed into approximation and detail components using wavelet transformation, with each component modeled using an ARIMA–GARCH structure. The forecasts were reconstructed using the inverse wavelet transformation. Results show that the proposed model consistently outperforms the conventional ARIMA–GARCH, achieving markedly lower error measures and higher explanatory power. Compared with earlier applications of wavelets, volatility hybrids in financial markets, this study is distinct in applying Coiflets wavelet to a macroeconomic indicator, highlighting their suitability for capturing both persistent trends and local shocks in CPI data. The findings demonstrate the practical value of the model as a reliable tool for policymakers and economists engaged in monitoring inflation, planning macroeconomics, and designing responsive economic strategies.